whatmodel.sh · find the best AI models you can run

Best AI models to run locally on a 16GB PC without a graphics card

A leaderboard of the best AI models you can run locally on a computer with 16GB of RAM and no graphics card. Spark-X2.5 4B is #1 right now, followed by MiMo-V2.6-Distill-Qwen-9B and MiniCPM5 2B. Models are ranked by one score that combines 472 published benchmarks. Each model shows the download that fits and the command to run it.

295 models · scores combine 472 benchmarks · updated how it works · find a leaderboard · benchmark your machine
task[all tasks][agent tasks][coding][chat & writing][math][vision & documents][web search]
Models for 16GB RAM · no GPU → ~10GB for models (CPU, slow) · local only · reset
+ more

Detect your exact chip or graphics card:

curl -G whatmodel.sh/ -d max=0 --data-urlencode "mac=$(sysctl -n hw.memsize machdep.cpu.brand_string)"Mac

curl -G whatmodel.sh/ -d max=0 --data-urlencode "ram=$(awk '/MemT/{print $2}' /proc/meminfo)" --data-urlencode "nv=$(nvidia-smi --query-gpu=name,memory.total --format=csv,noheader 2>/dev/null)"Linux

curl.exe -G whatmodel.sh/ -d max=0 --data-urlencode "ram=$((gcim Win32_ComputerSystem).TotalPhysicalMemory)" --data-urlencode "nv=$(nvidia-smi --query-gpu=name,memory.total --format=csv,noheader 2>$null)"Windows

$ terminal

detecting your OS…

Run curl "whatmodel.sh/?ram=16&gpu=none&max=0"to get this list in your terminal (add &format=json for agents).

ram[8][16][24][32][48][64][96][128][192][256]GB
gpu[apple silicon][graphics card][none]
price[local][$0.02][$0.1][$0.5][$2][$10][$50][any]max per M

detecting your OS…

Score vs memory to run

#modelscore↓where↓speed↓RAM↓context↓get itsources
#modelscore↓where↓speed↓RAM↓context↓get itsources
153Spark-X2.5 4B · 4-bit iFLYTEK (SparkLLM)66±1†local (slow)~4.2GB1Mllama.cpp › XHToken/Spark-X2.5-4B model card
161MiMo-V2.6-Distill-Qwen-9B · 4-bit Xiaomi65±2†local (slow)~8.6GBllama.cpp › XiaomiMiMo/MiMo-V2.6-Distill-Qwen-9B model card
186MiniCPM5 2B · 4-bit OpenBMB64±1†local (slow)~3.1GB131kllama.cpp › openbmb/MiniCPM5-2B model card
196Qwen3.5 9B · 4-bit Alibaba63±1local (slow)~7.4GB262kllama.cpp › Qwen/Qwen3.5-9B model cardXHToken/Spark-X2.5-4B model cardJetBrains/Mellum2.1-12B-A2.5B-Thinking model cardJetBrains/Mellum2-12B-A2.5B-Instruct model cardEpoch AI — internal evaluationsQwen/Qwen3.5-4B model card+17
202Gemma 4 12B · 4-bit Google62±1local (slow)~9.3GB262kllama.cpp › Gemma Team — Gemma 4 Technical ReportXHToken/Spark-X2.5-4B model cardgoogle/gemma-4-31B-it model card
226Qwen3.5 4B · 4-bit Alibaba61±1local (slow)~4.3GB262kllama.cpp › Qwen/Qwen3.5-9B model cardopenbmb/MiniCPM5-2B model cardXHToken/Spark-X2.5-4B model cardJetBrains/Mellum2-12B-A2.5B-Instruct model cardQwen/Qwen3.5-4B model cardLiquidAI/LFM2.5-2.6B model card+10
233Falcon-H1R 7B · 4-bit TII60±1†local (slow)~6.8GB262kllama.cpp › Falcon-H1R: Pushing the Reasoning Frontiers with a Hybrid Model for Efficient Test-Time Scaling
239Granite 4.2 8B · 4-bit IBM60±1local (slow)~7.1GB131kllama.cpp › ibm-granite/granite-4.2-30b model cardLMArena leaderboard dataset
247Gemma 2 9B IT SimPO · 4-bit Princeton NLP59±1local (slow)~7.6GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)WildBench v2 leaderboard data (Ai2)LMArena leaderboard datasetOpen Portuguese LLM Leaderboard (archived)
248Gemma 2 9B IT DPO (Princeton) · 4-bit Princeton NLP59±1local (slow)~7.6GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)WildBench v2 leaderboard data (Ai2)Open Portuguese LLM Leaderboard (archived)
253Nemotron-Cascade-8B · 4-bit NVIDIA59±1†local (slow)~6.8GBllama.cpp › NVIDIA — Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models
264Nemotron-Cascade-8B-Thinking · 4-bit NVIDIA58±2†local (slow)~6.8GBllama.cpp › NVIDIA — Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models
266Qwen3-VL 8B Thinking · 4-bit Alibaba58±1local (slow)~6.8GB256kllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)allenai/Olmo-3-7B-Think model cardChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)
275DeepSeek R1 0528 Qwen3 8B · 4-bit DeepSeek57±1local (slow)~6.8GB128kllama.cpp › Falcon-H1R: Pushing the Reasoning Frontiers with a Hybrid Model for Efficient Test-Time Scalingdeepseek-ai/DeepSeek-R1-0528-Qwen3-8B model cardEpoch AI — internal evaluationsBRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)EuroEval leaderboards (results collection), English generative leaderboard
276Granite 4.2 3B · 4-bit IBM57±1local (slow)~3.8GB131kllama.cpp › openbmb/MiniCPM5-2B model cardibm-granite/granite-4.2-30b model cardLMArena leaderboard dataset
281Qwen2.5-Omni 7B · 4-bit Alibaba57±2local (slow)~6.4GBllama.cpp › AHELM: A Holistic Evaluation of Audio-Language Models (Stanford CRFM, arXiv:2508.21376v2)
283QED-Nano · 4-bit LM Provers57±2local (slow)~4.1GB262kllama.cpp › MathArena ArXivMath / BrokenArXiv / ArXivLean model-output datasets
287Qwen3 4B Thinking 2507 · 4-bit Alibaba57±1local (slow)~4.1GB262kllama.cpp › Qwen/Qwen3-4B-Thinking-2507 model cardMathArena ArXivMath / BrokenArXiv / ArXivLean model-output datasetsMedmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks — authors' own runs of 61 modelsBRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)Qwen/Qwen3.5-0.8B model cardKnowledge Index of Noah's Ark (arXiv:2606.05104v2)+2
295Qwen3-VL 4B Thinking · 4-bit Alibaba56±1local (slow)~4.1GB256kllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)
302Qwen3 8B · 4-bit Alibaba56±1local (slow)~6.8GB33kllama.cpp › Falcon-H1R: Pushing the Reasoning Frontiers with a Hybrid Model for Efficient Test-Time ScalingQwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)Epoch AI — internal evaluationsMUA-RL: Multi-turn User-interacting Agent Reinforcement Learning for agentic tool use (arXiv:2508.18669v1)allenai/Olmo-3-7B-Think model cardSynthesize and Reward: Reinforcement Learning for Multi-Step Tool Use in Live Environments (arXiv:2606.03892v2)+14
305Nemotron Nano 9B v2 · 4-bit NVIDIA56±1local (slow)~7.6GB131kllama.cpp › Olmo 3NVIDIA — Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning ModelsNVIDIA vendor-reported benchmark scores for its own Nemotron models (Hugging Face model cards)allenai/Olmo-3-7B-Think model cardEuroEval leaderboards (results collection), English generative leaderboard
311Ministral 3 8B Reasoning 2512 · 4-bit Mistral55±1local (slow)~7GB262kllama.cpp › mistralai/Ministral-3-8B-Reasoning-2512 model cardmistralai/Ministral-3-3B-Reasoning-2512 model cardMedmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks — authors' own runs of 61 modelsSurgiQ: A Large-Scale Multi-Domain Benchmark for Evaluating Surgical Understanding in Large Language Models (arXiv:2606.08071v1)
314Molmo 2 8B · 4-bit Ai255±2local (slow)~6.8GBllama.cpp › Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (Ai2, arXiv 2601.10611v4)LMArena leaderboard dataset
317Qwen3 4B Instruct 2507 · 4-bit Alibaba55±1local (slow)~4.1GB262kllama.cpp › Qwen/Qwen3-4B-Instruct-2507 model cardMarco-MoE: Open Multilingual Mixture-of-Expert Language Models with Efficient UpcyclingQwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)Epoch AI — internal evaluationsRoblox LLM Leaderboard (Torpedo Software): RobloxQA v2.0 and Luau LeetcodeSynthesize and Reward: Reinforcement Learning for Multi-Step Tool Use in Live Environments (arXiv:2606.03892v2)+4
319Qwen3-VL 8B Instruct · 4-bit Alibaba55±1local (slow)~6.8GB256kllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)allenai/Olmo-3-7B-Instruct model cardChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
320Spark-X2.5 1.7B · 4-bit iFLYTEK (SparkLLM)55±1†local (slow)~2.7GB1Mllama.cpp › XHToken/Spark-X2.5-4B model card
325Bielik 11B v3.0 Instruct · 4-bit SpeakLeash55±2local (slow)~8.5GBllama.cpp › Reassessing High-Performing LLMs on Polish Medical Exams: True Competence or Bias-Driven Performance? (arXiv:2606.12250v1)SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardEuroEval leaderboards (results collection), English generative leaderboard
329EXAONE Deep 7.8B · 4-bit LG AI Research55±2†local (slow)~6.5GB33kllama.cpp › LG AI Research vendor-reported benchmark scores for its own EXAONE models (Hugging Face model cards)
335Nemotron 3 Nano 4B · 4-bit NVIDIA54±1local (slow)~4.4GB262kllama.cpp › openbmb/MiniCPM5-2B model cardNVIDIA vendor-reported benchmark scores for its own Nemotron models (Hugging Face model cards)
336Granite 4.1 8B · 4-bit IBM54±1local (slow)~7.1GB131kOllama › ibm-granite/granite-4.1-8b model cardLMArena leaderboard datasetSynthesize and Reward: Reinforcement Learning for Multi-Step Tool Use in Live Environments (arXiv:2606.03892v2)
338Bielik 11B v2.6 Instruct · 4-bit SpeakLeash54±3local (slow)~8.5GBllama.cpp › SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardEuroEval leaderboards (results collection), English generative leaderboard
349GLM-4 9B Chat 1M · 4-bit Zhipu AI (THUDM)53±5local (slow)~8.1GB1Mllama.cpp › NVIDIA RULER repository README results tableOpen Portuguese LLM Leaderboard (archived)
361Olmo 3 7B Think · 4-bit Ai253±1local (slow)~6.2GB66kllama.cpp › Olmo 3Medmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks — authors' own runs of 61 modelsallenai/Olmo-3-7B-Think model cardJetBrains/Mellum2-12B-A2.5B-Thinking model cardEuroEval leaderboards (results collection), English generative leaderboard
366Gemma 4 E4B · 4-bit Google53±1local (slow)~7.1GB131kllama.cpp › Gemma Team — Gemma 4 Technical Reportopenbmb/MiniCPM5-2B model cardJetBrains/Mellum2.1-12B-A2.5B-Thinking model cardXHToken/Spark-X2.5-4B model cardLiquidAI/LFM2.5-2.6B model cardEuroEval leaderboards (results collection), English generative leaderboard+7
367Ministral 3 8B · 4-bit Mistral53±1local (slow)~7GB262kllama.cpp › Marco-MoE: Open Multilingual Mixture-of-Expert Language Models with Efficient UpcyclingMedmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks — authors' own runs of 61 modelsEuroEval leaderboards (results collection), English generative leaderboardmistralai/Ministral-3-14B-Instruct-2512 model cardContractBench: Can LLM Agents Preserve Observation Contracts? (arXiv:2605.17281v2)mistralai/Ministral-3-8B-Instruct-2512 model card+2
368Qwen3-VL 4B Instruct · 4-bit Alibaba53±1local (slow)~4.1GB256kllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
370LFM2.5 2.6B · 4-bit Liquid AI53±1local (slow)~3.3GB131kllama.cpp › openbmb/MiniCPM5-2B model cardLiquidAI/LFM2.5-2.6B model card
371Bielik 11B v2.1 Instruct · 4-bit SpeakLeash53±3local (slow)~8.5GBllama.cpp › SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)
374Qwen2.5-Math 7B Instruct · 4-bit Alibaba52±2†local (slow)~6.4GB4kllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)
375LFM2.5 8B-A1B · 4-bit Liquid AI52±1local (slow)~7GB128kllama.cpp › openbmb/MiniCPM5-2B model cardLiquidAI/LFM2.5-8B-A1B model card
379Bielik 11B v2.2 Instruct · 4-bit SpeakLeash52±3local (slow)~8.5GBllama.cpp › SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)
385Gemma 3 12B · 4-bit Google52±1local (slow)~9.2GB131kllama.cpp › Google DeepMind Gemma 3 270M IT model cardMarco-MoE: Open Multilingual Mixture-of-Expert Language Models with Efficient UpcyclingMedmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks — authors' own runs of 61 modelsGemma Team — Gemma 3 Technical ReportEpoch AI — internal evaluationsMIMIC Clinical Decision Making leaderboard+10
386Qwen2.5 Coder 7B Instruct · 4-bit Alibaba52±1local (slow)~6.4GB131kllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)LiveCodeBench official website results (2024-12-26)Aider code editing benchmark runs (old edit leaderboard)CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)Open Portuguese LLM Leaderboard (archived)
388Cogito v1 Preview Llama 8B · 4-bit Deep Cogito51±5local (slow)~6.6GB128kllama.cpp › EuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
391OpenCoder-8B-Instruct · 4-bit INF (infly)51±1local (slow)~6.4GBllama.cpp › EvalPlus / EvalPerf published evaluation runsEuroEval leaderboards (results collection), English generative leaderboard
394Ministral 3 3B Reasoning 2512 · 4-bit Mistral51±1local (slow)~3.8GB262kllama.cpp › mistralai/Ministral-3-8B-Reasoning-2512 model cardmistralai/Ministral-3-3B-Reasoning-2512 model cardMedmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks — authors' own runs of 61 modelsEuroEval leaderboards (results collection), English generative leaderboard
395Qwen3 4B · 4-bit Alibaba51±1local (slow)~4.1GB33kllama.cpp › Qwen/Qwen3-4B-Thinking-2507 model cardQwen/Qwen3-4B-Instruct-2507 model cardLFM2 Technical Report — Liquid AI own SmolLM3 evaluationHuggingFaceTB/SmolLM3-3B model card (Evaluation section)Epoch AI — internal evaluationsArena-Hard-Auto v2.0-Preview leaderboard (repository README)+8
397Llama-3-Instruct-8B-SimPO-ExPO · 4-bit Chujie Zheng (ExPO)51±1local (slow)~6.6GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)WildBench v2 leaderboard data (Ai2)Open Portuguese LLM Leaderboard (archived)
400Llama-3-Instruct-8B-SimPO · 4-bit Princeton NLP51±1local (slow)~6.6GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)WildBench v2 leaderboard data (Ai2)Open Portuguese LLM Leaderboard (archived)
403Bielik 11B v2.0 Instruct · 4-bit SpeakLeash51±3local (slow)~8.5GBllama.cpp › SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)
404OpenThinker3 7B · 4-bit Open Thoughts50±1local (slow)~6.4GB33kllama.cpp › Olmo 3allenai/Olmo-3-7B-Think model cardBRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)
405CodeQwen1.5 7B Chat · 4-bit Alibaba50±1local (slow)~6.4GB64kllama.cpp › EvalPlus / EvalPerf published evaluation runsQwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)BigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4
409EXAONE Deep 2.4B · 4-bit LG AI Research50±2†local (slow)~3.2GB33kOllama › LG AI Research vendor-reported benchmark scores for its own EXAONE models (Hugging Face model cards)
416Bielik 11B v2.3 Instruct · 4-bit SpeakLeash50±2local (slow)~8.6GBllama.cpp › SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)Berkeley Function Calling Leaderboard result archiveSpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)EuroEval leaderboards (results collection), English generative leaderboard
417Qwen2.5-VL 7B Instruct · 4-bit Alibaba50±1local (slow)~6.4GBllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)
428OpenThinker2 7B · 4-bit Open Thoughts49±2†local (slow)~6.4GB33kllama.cpp › open-thoughts/OpenThinker2-7B model card
430Granite 4.1 3B · 4-bit IBM49±1†local (slow)~3.7GBOllama › ibm-granite/granite-4.1-8b model card
431OpenCodeInterpreter-DS-6.7B · 4-bit M-A-P49±1local (slow)~5.8GBllama.cpp › EvalPlus / EvalPerf published evaluation runs
432Artigenz-Coder-DS-6.7B · 4-bit Artigenz49±1local (slow)~5.8GBllama.cpp › EvalPlus / EvalPerf published evaluation runs
433LFM2.5 1.2B Thinking · 4-bit Liquid AI49±2local (slow)~2.2GB33kllama.cpp › LiquidAI/LFM2.5-1.2B-Thinking model cardEuroEval leaderboards (results collection), English generative leaderboard
434Qwen3-VL 2B Thinking · 4-bit Alibaba49±1local (slow)~2.7GB256kllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)
436Pixtral 12B (2409) · 4-bit Mistral49±2local (slow)~9.6GBllama.cpp › Pixtral 12B (Mistral AI, arXiv 2410.07073v2)LMArena leaderboard datasetVisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback? (arXiv:2608.10408v1)Ai2 Molmo-72B-0924 model card: Human Preference Elo Rating column
438MegaBeam-Mistral-7B-512k · 4-bit AWS (aws-prototyping)49±4local (slow)~6.1GBllama.cpp › NVIDIA RULER repository README results tableOpen Portuguese LLM Leaderboard (archived)
441AceReason-Nemotron-1.1-7B · 4-bit NVIDIA49±2†local (slow)~6.4GBllama.cpp › NVIDIA — Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models
443Yi-Coder 9B Chat · 4-bit 01.AI49±1local (slow)~7.1GB128kllama.cpp › LiveCodeBench official website results (2024-12-26)Aider code editing benchmark runs (old edit leaderboard)CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)EvalPlus / EvalPerf published evaluation runsOpen Portuguese LLM Leaderboard (archived)
444Qwen2-Math 7B Instruct · 4-bit Alibaba49±2†local (slow)~6.1GB4kOllama › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)
445Magicoder-S-DS 6.7B · 4-bit UIUC (ise-uiuc)49±1local (slow)~5.8GBllama.cpp › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4CRUXEval published evaluation results (authors' generations and scores)
446InternLM2.5 7B Chat · 4-bit Shanghai AI Laboratory (InternLM)49±2local (slow)~6.4GBllama.cpp › InternLM GitHub repository model cards (Shanghai AI Laboratory)Open Portuguese LLM Leaderboard (archived)
448Qwen2.5 7B Instruct · 4-bit Alibaba48±1local (slow)~6.4GB131kllama.cpp › microsoft/Phi-4-mini-instruct model cardWildEval ZeroEval published results2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsLiveCodeBench official website results (2024-12-26)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)Epoch AI — internal evaluations+12
449DeepSeek R1 Distill Qwen 7B · 4-bit DeepSeek48±1local (slow)~6.4GB131kllama.cpp › open-thoughts/OpenThinker2-7B model carddeepseek-ai/DeepSeek-R1 model cardallenai/Olmo-3-7B-Think model cardMicrosoft — Phi-4-Mini Technical Reportmicrosoft/Phi-4-mini-reasoning model cardopen-thoughts/OpenThinker-7B model card+6
455T5 Large · 4-bit Google48±9local (slow)~4.6GB1khow to run › COPEN — EMNLP 2022 original zero-shot evaluations
456WaveCoder-Ultra-6.7B · 4-bit Microsoft48±1local (slow)~5.8GBllama.cpp › EvalPlus / EvalPerf published evaluation runs
460Gemma 3n E4B · 4-bit Google48±1local (slow)~6.2GB32kllama.cpp › Google DeepMind Gemma 3n E2B IT model cardLMArena leaderboard datasetOpen Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)EuroEval leaderboards (results collection), English generative leaderboard
462Granite 4.0 H Tiny (7B-A1B) · 4-bit IBM48±1local (slow)~5.9GB131kOllama › ibm-granite/granite-4.0-h-small model cardEuroEval leaderboards (results collection), English generative leaderboardVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
463DeepSeek-Coder 7B Instruct v1.5 · 4-bit DeepSeek48±1local (slow)~5.9GBllama.cpp › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsCodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)
465Qwen2.5 Coder 3B Instruct · 4-bit Alibaba48±1local (slow)~3.5GB33kOllama › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)Aider code editing benchmark runs (old edit leaderboard)
466DeepSeek-Coder 6.7B Instruct · 4-bit DeepSeek48±1local (slow)~5.8GB16kllama.cpp › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4LiveCodeBench official website results (2024-12-26)CRUXEval published evaluation results (authors' generations and scores)
467MiniCPM3-4B · 4-bit OpenBMB48±2local (slow)~4.1GB32kllama.cpp › LiveCodeBench official website results (2024-12-26)Berkeley Function Calling Leaderboard result archive
468Bielik 4.5B v3.0 Instruct · 4-bit SpeakLeash48±2local (slow)~4.5GBllama.cpp › SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)Reassessing High-Performing LLMs on Polish Medical Exams: True Competence or Bias-Driven Performance? (arXiv:2606.12250v1)SpeakLeash Polish EQ-Bench leaderboardEuroEval leaderboards (results collection), English generative leaderboard
469Qwen3.5 2B · 4-bit Alibaba47±1local (slow)~2.9GB262kllama.cpp › Qwen/Qwen3.5-2B model cardopenbmb/MiniCPM5-2B model cardXHToken/Spark-X2.5-4B model cardCohereLabs/North-Micro-Vision-Instruct model card (benchmark table)EuroEval leaderboards (results collection), English generative leaderboardMathArena ArXivMath / BrokenArXiv / ArXivLean model-output datasets+7
470Gemma 2 9B IT · 4-bit Google47±1local (slow)~7.6GB8kllama.cpp › microsoft/Phi-4-mini-instruct model cardWildEval ZeroEval published results2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsEpoch AI — internal evaluationsInternLM GitHub repository model cards (Shanghai AI Laboratory)+14
472GLM-4 9B Chat · 4-bit Zhipu AI (Z.ai)47±2local (slow)~7.9GB128kOllama › WildBench v2 leaderboard data (Ai2)SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardBerkeley Function Calling Leaderboard result archiveOpen Portuguese LLM Leaderboard (archived)ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools (Team GLM, arXiv 2406.12793v2)+1
474Phi-4-mini-reasoning 3.8B · 4-bit Microsoft47±2local (slow)~4.1GB128kllama.cpp › microsoft/Phi-4-mini-reasoning model cardBRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)EuroEval leaderboards (results collection), English generative leaderboard
477DeepSeek R1 Distill Llama 8B · 4-bit DeepSeek47±1local (slow)~6.6GB131kllama.cpp › deepseek-ai/DeepSeek-R1 model cardMicrosoft — Phi-4-Mini Technical Reportmicrosoft/Phi-4-mini-reasoning model cardCodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)BRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)EuroEval leaderboards (results collection), English generative leaderboard+1
478EXAONE 4.0 1.2B · 4-bit LG AI Research47±2†local (slow)~2.8GB66kllama.cpp › LG AI Research vendor-reported benchmark scores for its own EXAONE models (Hugging Face model cards)
480speechless-coder-ds-6.7b · 4-bit uukuguy47±1local (slow)~5.8GBllama.cpp › EvalPlus / EvalPerf published evaluation runs
481Gemma 4 E2B · 4-bit Google47±1local (slow)~5.1GB131kllama.cpp › Gemma Team — Gemma 4 Technical Reportopenbmb/MiniCPM5-2B model cardXHToken/Spark-X2.5-4B model cardCohereLabs/North-Micro-Vision-Instruct model card (benchmark table)LiquidAI/LFM2.5-2.6B model cardEuroEval leaderboards (results collection), English generative leaderboard+6
482Magicoder-S-CL-7B · 4-bit UIUC (ise-uiuc)47±1local (slow)~5.8GBllama.cpp › EvalPlus / EvalPerf published evaluation runs
483Aya Expanse 8B · 4-bit Cohere47±2local (slow)~6.9GB8kllama.cpp › LMArena leaderboard datasetInception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)Open Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)Open Portuguese LLM Leaderboard (archived)EuroEval leaderboards (results collection), English generative leaderboardVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
486InternLM2.5 7B Chat 1M · 4-bit Shanghai AI Laboratory (InternLM)47±4local (slow)~6.4GBllama.cpp › NVIDIA RULER repository README results tableOpen Portuguese LLM Leaderboard (archived)
487EuroLLM 9B Instruct · 4-bit UTTER (EuroLLM)47±3local (slow)~7.4GB4kllama.cpp › SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)EuroEval leaderboards (results collection), English generative leaderboard
488Granite 4.0 H Micro 3B · 4-bit IBM47±1local (slow)~3.5GB131kOllama › ibm-granite/granite-4.0-h-small model cardEuroEval leaderboards (results collection), English generative leaderboardVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
492MedGemma 4B IT · 4-bit Google46±3local (slow)~4.1GBllama.cpp › Reassessing High-Performing LLMs on Polish Medical Exams: True Competence or Bias-Driven Performance? (arXiv:2606.12250v1)BRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)Open Portuguese LLM Leaderboard (archived)SurgiQ: A Large-Scale Multi-Domain Benchmark for Evaluating Surgical Understanding in Large Language Models (arXiv:2606.08071v1)
494Falcon-H1 7B Instruct · 4-bit TII46±3local (slow)~6.3GB262kllama.cpp › Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)Open Portuguese LLM Leaderboard (archived)
496Cogito v1 Preview Llama 3B · 4-bit Deep Cogito46±4local (slow)~3.8GB128kllama.cpp › EuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
498Hermes 2 Pro Llama 3 8B · 4-bit Nous Research46±3local (slow)~6.6GB8kllama.cpp › SpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)Berkeley Function Calling Leaderboard result archive
500Granite 4.0 Micro 3B · 4-bit IBM46±1local (slow)~3.7GB131kllama.cpp › ibm-granite/granite-4.0-h-small model cardEpoch AI — internal evaluationsEuroEval leaderboards (results collection), English generative leaderboardVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
504T5 Base · 4-bit Google46±9local (slow)~2.4GB1khow to run › COPEN — EMNLP 2022 original zero-shot evaluations
505OpenChat 3.5 0106 · 4-bit OpenChat46±1local (slow)~6.1GB8kllama.cpp › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboard+1
506Qwen2-VL 7B Instruct · 4-bit Alibaba46±1local (slow)~7.2GBllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)Pixtral 12B (Mistral AI, arXiv 2410.07073v2)LMArena leaderboard datasetCapArena: Benchmarking and Analyzing Detailed Image Captioning in the LLM Era (arXiv:2503.12329v1)Ai2 Molmo-72B-0924 model card: Human Preference Elo Rating columnVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
507Qwen2.5-Math 1.5B Instruct · 4-bit Alibaba46±2†local (slow)~2.5GB4kllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)
515Granite 3.3 8B Instruct · 4-bit IBM46±1local (slow)~6.6GB128kllama.cpp › ibm-granite/granite-3.3-8b-instruct model cardallenai/Olmo-3-7B-Instruct model cardDiagnosticIQ: A Benchmark for LLM-Based Industrial Maintenance Action Recommendation from Symbolic Rules (arXiv:2605.08614v1)Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
518OpenChat 3.5 0106 Gemma · 4-bit OpenChat46±3local (slow)~6.9GB8kllama.cpp › SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)
519Olmo 3 7B Instruct · 4-bit Ai246±1local (slow)~6.2GB66kllama.cpp › Olmo 3JetBrains/Mellum2-12B-A2.5B-Instruct model cardMedmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks — authors' own runs of 61 modelsallenai/Olmo-3-7B-Instruct model cardEuroEval leaderboards (results collection), English generative leaderboard
524Ministral 3 3B · 4-bit Mistral45±1local (slow)~3.7GB262kllama.cpp › Marco-MoE: Open Multilingual Mixture-of-Expert Language Models with Efficient UpcyclingCohereLabs/North-Micro-Vision-Instruct model card (benchmark table)Medmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks — authors' own runs of 61 modelsmistralai/Ministral-3-14B-Instruct-2512 model cardmistralai/Ministral-3-8B-Instruct-2512 model cardEuroEval leaderboards (results collection), English generative leaderboard+2
526OpenThinker 7B · 4-bit Open Thoughts45±2†local (slow)~6.4GB33kllama.cpp › open-thoughts/OpenThinker2-7B model cardopen-thoughts/OpenThinker-7B model card
534Qwen2-Math 1.5B Instruct · 4-bit Alibaba45±2†local (slow)~2.5GB4kOllama › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)
536Llama 3.1 8B Instruct · 4-bit Meta45±0local (slow)~6.6GB128kllama.cpp › EvalPlus / EvalPerf published evaluation runsIBM Granite 3.0 Language Models technical report (GitHub)microsoft/Phi-4-mini-instruct model cardWildEval ZeroEval published results2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasets+25
538Mistral NeMo Instruct (2407) · 4-bit Mistral44±1local (slow)~9.4GB128kllama.cpp › WildEval ZeroEval published resultsLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsEpoch AI — internal evaluationsWildBench v2 leaderboard data (Ai2)EvalPlus / EvalPerf published evaluation runsSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)+9
540Phi-3-mini-4k-instruct (June 2024 update) · 4-bit Microsoft44±1local (slow)~4GBllama.cpp › WildEval ZeroEval published resultsLMArena leaderboard datasetEuroEval leaderboards (results collection), English generative leaderboard
541Yi 1.5 9B Chat · 4-bit 01.AI44±1local (slow)~7.1GB4kllama.cpp › WildEval ZeroEval published resultsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4WildBench v2 leaderboard data (Ai2)CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)SpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)+1
542Command R7B (12-2024) · 4-bit Cohere44±1local (slow)~6.9GB128kllama.cpp › Cohere — Command A technical reportInception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)Berkeley Function Calling Leaderboard result archiveOpen Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)
543Qwen2-Audio 7B Instruct · 4-bit Alibaba44±3local (slow)~6.5GBllama.cpp › AHELM: A Holistic Evaluation of Audio-Language Models (Stanford CRFM, arXiv:2508.21376v2)
544Gemma 3 4B · 4-bit Google44±1local (slow)~4.1GB131kllama.cpp › Google DeepMind Gemma 3 270M IT model cardLFM2 Technical Report — Liquid AI own SmolLM3 evaluationMedmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks — authors' own runs of 61 modelsGemma Team — Gemma 3 Technical ReportMIMIC Clinical Decision Making leaderboardLMArena leaderboard dataset+11
545Eurus-2-7B-PRIME · 4-bit PRIME-RL44±2local (slow)~6.4GB4kllama.cpp › PRIME-RL Eurus-2-7B-PRIME model card, Evaluation table (own models)Epoch AI — internal evaluations
546Llama 3.1 Tulu 3 8B · 4-bit Ai244±1local (slow)~6.6GB131kllama.cpp › microsoft/Phi-4-mini-instruct model card2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsLMArena leaderboard datasetEuroEval leaderboards (results collection), English generative leaderboard
547Qwen2.5-VL 3B Instruct · 4-bit Alibaba44±1local (slow)~3.5GBllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)
550Qwen2 7B Instruct · 4-bit Alibaba44±1local (slow)~6.4GB131kllama.cpp › WildEval ZeroEval published resultsInternLM GitHub repository model cards (Shanghai AI Laboratory)LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsQwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)BigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)+4
551Falcon3 7B Instruct · 4-bit TII44±3local (slow)~6.3GB32kllama.cpp › CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)Berkeley Function Calling Leaderboard result archiveOpen Portuguese LLM Leaderboard (archived)
553Qwen3-VL 2B Instruct · 4-bit Alibaba44±1local (slow)~2.7GB256kllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)CohereLabs/North-Micro-Vision-Instruct model card (benchmark table)ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
555Granite 8B Code Instruct (4K) · 4-bit IBM44±1local (slow)~6.6GBllama.cpp › Granite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4EuroEval leaderboards (results collection), English generative leaderboard
556Falcon3 10B Instruct · 4-bit TII44±3local (slow)~8.1GB32kllama.cpp › CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)Berkeley Function Calling Leaderboard result archiveOpen Portuguese LLM Leaderboard (archived)
558Qwen3 1.7B · 4-bit Alibaba44±1local (slow)~2.9GB33kllama.cpp › Marco-MoE: Open Multilingual Mixture-of-Expert Language Models with Efficient UpcyclingQwen/Qwen3.5-2B model cardQwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)LiquidAI/LFM2-1.2B model cardHuggingFaceTB/SmolLM3-3B model card (Evaluation section)LiquidAI/LFM2.5-1.2B-Thinking model card+10
559Phi-4-mini 3.8B · 4-bit Microsoft44±1local (slow)~4.1GB128kllama.cpp › microsoft/Phi-4-mini-instruct model cardMicrosoft — Phi-4-Mini Technical ReportDaekeun Kim Korean LLM benchmark runsSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardKorean Canonical Legal Benchmark: Toward Knowledge-Independent Evaluation of LLMs’ Legal Reasoning Capabilities+4
560Ministral 8B (2410) · 4-bit Mistral44±1local (slow)~6.6GB128kllama.cpp › microsoft/Phi-4-mini-instruct model card2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsEpoch AI — internal evaluationsMistral AI vendor-reported benchmark scores for its own models (Hugging Face model cards)LMArena leaderboard datasetBRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)+6
561Granite 3.2 8B Instruct · 4-bit IBM44±1local (slow)~6.6GB131kllama.cpp › ibm-granite/granite-3.3-8b-instruct model cardBerkeley Function Calling Leaderboard result archiveVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
563Llama 3.1 8B (base) · 4-bit Meta44±3local (slow)~6.6GBllama.cpp › Open Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)EuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
564LFM2.5 1.2B Instruct · 4-bit Liquid AI44±2local (slow)~2.3GB33kllama.cpp › LiquidAI/LFM2.5-1.2B-Thinking model cardEuroEval leaderboards (results collection), English generative leaderboard
565Granite 3.0 8B Instruct · 4-bit IBM43±1local (slow)~6.6GB4kllama.cpp › IBM Granite 3.0 Language Models technical report (GitHub)LMArena leaderboard datasetCodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)EuroEval leaderboards (results collection), English generative leaderboardVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
566Phi-3.5-mini-instruct · 4-bit Microsoft43±1local (slow)~4GB128kllama.cpp › microsoft/Phi-4-mini-instruct model cardWildEval ZeroEval published resultsLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsDaekeun Kim Korean LLM benchmark runsMicrosoft — Phi-4-Mini Technical ReportSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)+4
567Hermes 2 Theta Llama 3 8B · 4-bit Nous Research43±2local (slow)~6.6GB8kllama.cpp › WildBench v2 leaderboard data (Ai2)SpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)
571Gemma 3n E2B IT · 4-bit Google43±2local (slow)~4.7GB32kllama.cpp › Google DeepMind Gemma 3n E2B IT model cardOpen Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)EuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
573T5 Small · 4-bit Google43±9local (slow)~1.8GB1khow to run › COPEN — EMNLP 2022 original zero-shot evaluations
574MiniCPM-V 2.6 · 4-bit OpenBMB43±2local (slow)~6.4GBllama.cpp › CharXiv leaderboard (princeton-nlp/CharXiv README)LMArena leaderboard datasetCapArena: Benchmarking and Analyzing Detailed Image Captioning in the LLM Era (arXiv:2503.12329v1)PhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understanding (arXiv:2501.16411v2)
575Qwen2.5 Coder 1.5B Instruct · 4-bit Alibaba43±1local (slow)~2.5GB33kOllama › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)Aider code editing benchmark runs (old edit leaderboard)
576LFM2 2.6B · 4-bit Liquid AI43±2local (slow)~3.1GB33kllama.cpp › LFM2 Technical Report — Liquid AI own SmolLM3 evaluationEuroEval leaderboards (results collection), English generative leaderboard
577SmolLM3 3B · 4-bit Hugging Face43±1local (slow)~4GB66kllama.cpp › LFM2 Technical Report — Liquid AI own SmolLM3 evaluationHuggingFaceTB/SmolLM3-3B model card (Evaluation section)Medmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks — authors' own runs of 61 modelsInception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)EuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
578Granite 3.1 8B Instruct · 4-bit IBM43±1local (slow)~6.6GB128kllama.cpp › ibm-granite/granite-3.3-8b-instruct model cardibm-granite/granite-3.1-3b-a800m-instruct model cardLMArena leaderboard datasetBerkeley Function Calling Leaderboard result archiveOpen Portuguese LLM Leaderboard (archived)Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
579Llama 3.2 11B Vision Instruct · 4-bit Meta43±2local (slow)~8.8GB128kOllama › Pixtral 12B (Mistral AI, arXiv 2410.07073v2)LMArena leaderboard datasetInception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)Open Portuguese LLM Leaderboard (archived)
580Qwen2.5 3B Instruct · 4-bit Alibaba43±1local (slow)~3.7GB33kllama.cpp › microsoft/Phi-4-mini-instruct model cardWildEval ZeroEval published resultsSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)SpeakLeash Polish EQ-Bench leaderboard+5
581DeepSeek-Coder 1.3B Instruct · 4-bit DeepSeek43±1local (slow)~2.4GB16kOllama › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4LiveCodeBench official website results (2024-12-26)CRUXEval published evaluation results (authors' generations and scores)CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)
582Aya 23 8B · 4-bit Cohere For AI43±3local (slow)~6.9GB8kllama.cpp › SpeakLeash Polish EQ-Bench leaderboardOpen Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)Open Portuguese LLM Leaderboard (archived)EuroEval leaderboards (results collection), English generative leaderboard
584CodeGemma 7B IT (v1.0) · 4-bit Google43±1local (slow)~7.1GBllama.cpp › EvalPlus / EvalPerf published evaluation runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4CodeGemma: Open Code Models Based on Gemma — authors' own runs
587Mathstral 7B v0.1 · 4-bit Mistral AI42±1local (slow)~6.1GB32kllama.cpp › LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsMMLU-Pro author-released evaluation runsWildEval ZeroEval published resultsMistral AI vendor-reported benchmark scores for its own models (Hugging Face model cards)
588Yi-1.5 9B (base) · 4-bit 01.AI42±2local (slow)~7.1GB4kllama.cpp › Open Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)BigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4EuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
590Llama 3 8B Instruct · 4-bit Meta42±0local (slow)~6.6GB8kllama.cpp › EvalPlus / EvalPerf published evaluation runsWildEval ZeroEval published resultsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsInternLM GitHub repository model cards (Shanghai AI Laboratory)LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)+19
593Hermes 3 Llama 3.1 8B · 4-bit Nous Research42±1local (slow)~6.6GB128kllama.cpp › Nous Research vendor-reported scores for its own Hermes models (Hermes 3 Technical Report)Aider code editing benchmark runs (old edit leaderboard)Open Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)Open Portuguese LLM Leaderboard (archived)
594Falcon2 11B (base) · 4-bit TII42±3local (slow)~8.2GB8kOllama › SpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)EuroEval leaderboards (results collection), English generative leaderboard
596speechless-starcoder2-7b · 4-bit uukuguy42±1local (slow)~6.1GBllama.cpp › EvalPlus / EvalPerf published evaluation runs
599OLMo 2 7B Instruct (1124) · 4-bit Allen Institute for AI42±1local (slow)~6.2GB4kllama.cpp › 2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsallenai/Olmo-3-7B-Instruct model cardEuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
603OpenChat 3.5 · 4-bit OpenChat41±2local (slow)~6.1GB8kllama.cpp › OpenChat 3.5 model card (openchat/openchat_3.5)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetEQ-Bench: An Emotional Intelligence Benchmark for Large Language Models — authors' own runs
614CodeQwen1.5 7B (base) · 4-bit Alibaba41±1local (slow)~6.4GB66kllama.cpp › EvalPlus / EvalPerf published evaluation runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4
615Granite 4.0 1B · 4-bit IBM41±1local (slow)~2.6GB128kllama.cpp › ibm-granite/granite-4.0-h-1b model cardLiquidAI/LFM2.5-1.2B-Thinking model cardEpoch AI — internal evaluations
619Granite 4.0 H 1B · 4-bit IBM40±1local (slow)~2.4GB131kllama.cpp › ibm-granite/granite-4.0-h-1b model cardLiquidAI/LFM2.5-1.2B-Thinking model card
620Llama3-OpenBioLLM-8B · 4-bit Saama AI Labs40±2local (slow)~6.6GBllama.cpp › Open Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)BRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)
621Nous Hermes 2 Mistral 7B DPO · 4-bit Nous Research40±2local (slow)~6.1GB33kllama.cpp › Open Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)Open Portuguese LLM Leaderboard (archived)
626Starling-LM 7B beta · 4-bit Nexusflow40±1local (slow)~6.1GB8kllama.cpp › LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)WildBench v2 leaderboard data (Ai2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard dataset+3
627LFM2 1.2B · 4-bit Liquid AI40±2local (slow)~2.3GB33kllama.cpp › LiquidAI/LFM2-1.2B model cardEuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
630Bielik 1.5B v3.0 Instruct · 4-bit SpeakLeash40±3local (slow)~2.6GBllama.cpp › SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)EuroEval leaderboards (results collection), English generative leaderboard
631OpenHermes 2.5 Mistral 7B · 4-bit Nous Research39±1local (slow)~6.1GBllama.cpp › LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsOpenChat 3.5 model card (openchat/openchat_3.5)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Hallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard dataset+2
632Granite 3.3 2B Instruct · 4-bit IBM39±1†local (slow)~3.1GB128kllama.cpp › ibm-granite/granite-3.3-8b-instruct model card
633Phi-3-mini-128k-instruct (Apr 2024) · 16-bit Microsoft39±2local (slow)~9.5GBhow to run › Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetCodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)Open Portuguese LLM Leaderboard (archived)EuroEval leaderboards (results collection), English generative leaderboard+1
634Mistral 7B Instruct v0.3 · 4-bit Mistral39±1local (slow)~6.1GB32kllama.cpp › LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Mistral AI vendor-reported benchmark scores for its own models (Hugging Face model cards)SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardEpoch AI — internal evaluations+4
636DeepSeek-Coder 6.7B Base · 4-bit DeepSeek39±1local (slow)~5.8GB16kllama.cpp › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsStarCoder 2 and The Stack v2: The Next Generation — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4CRUXEval published evaluation results (authors' generations and scores)
637Starling-LM 7B alpha · 4-bit UC Berkeley (NEST)39±1local (slow)~6.1GB8kllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Hallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)+2
638Granite 8B Code Base (4K) · 4-bit IBM39±1local (slow)~6.6GBllama.cpp › Granite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4EuroEval leaderboards (results collection), English generative leaderboard
641Falcon-H1 3B Instruct · 4-bit TII39±4local (slow)~3.5GB131kllama.cpp › Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)Open Portuguese LLM Leaderboard (archived)
642Phi-2 (2.7B) · 4-bit Microsoft39±1local (slow)~3.2GB2kOllama › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsMiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies (Hu et al.)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)CRUXEval published evaluation results (authors' generations and scores)Open Portuguese LLM Leaderboard (archived)+2
643Granite 3.0 2B Instruct · 4-bit IBM39±1local (slow)~3.2GB4kllama.cpp › IBM Granite 3.0 Language Models technical report (GitHub)LMArena leaderboard datasetCodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)EuroEval leaderboards (results collection), English generative leaderboardVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
646MistralHermes-CodePro-7B-v1 · 4-bit beowolx39±1local (slow)~6.1GBllama.cpp › EvalPlus / EvalPerf published evaluation runs
647Llama-3-Instruct-8B-SimPO-v0.2 · 4-bit Princeton NLP39±1local (slow)~6.6GBllama.cpp › WildEval ZeroEval published resultsWildBench v2 leaderboard data (Ai2)Open Portuguese LLM Leaderboard (archived)
648Llama 3.2 3B Instruct · 4-bit Meta38±1local (slow)~3.6GB131kllama.cpp › microsoft/Phi-4-mini-instruct model cardIBM Granite 3.0 Language Models technical report (GitHub)LFM2 Technical Report — Liquid AI own SmolLM3 evaluationAider code editing benchmark runs (old edit leaderboard)microsoft/Phi-4-mini-reasoning model cardMicrosoft — Phi-4-Mini Technical Report+11
651CodeGemma 7B (v1.0) · 4-bit Google38±1local (slow)~7.1GBllama.cpp › EvalPlus / EvalPerf published evaluation runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsCodeGemma: Open Code Models Based on Gemma — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4
654Phi-3-mini-4k-instruct (Apr 2024) · 4-bit Microsoft38±2local (slow)~4GBllama.cpp › MMLU-Pro author-released evaluation runsPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetSpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)
655Apertus 8B Instruct (2509) · 4-bit Swiss AI38±4local (slow)~6.9GB66kllama.cpp › allenai/Olmo-3-7B-Instruct model cardEuroEval leaderboards (results collection), English generative leaderboardMiðeind Icelandic LLM Leaderboard
657Granite 3B Code Instruct (2K) · 4-bit IBM38±2local (slow)~3.7GBllama.cpp › Granite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4EuroEval leaderboards (results collection), English generative leaderboard
659BioMistral-7B · 4-bit BioMistral38±3local (slow)~6.1GBllama.cpp › BRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)Reassessing High-Performing LLMs on Polish Medical Exams: True Competence or Bias-Driven Performance? (arXiv:2606.12250v1)SurgiQ: A Large-Scale Multi-Domain Benchmark for Evaluating Surgical Understanding in Large Language Models (arXiv:2606.08071v1)
660Qwen3.5 0.8B · 4-bit Alibaba38±1local (slow)~2GB262kllama.cpp › Qwen/Qwen3.5-2B model cardEuroEval leaderboards (results collection), English generative leaderboardQwen/Qwen3.5-0.8B model cardChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)Knowledge Index of Noah's Ark (arXiv:2606.05104v2)BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)+2
662SOLAR 10.7B Instruct v1.0 · 4-bit Upstage37±1local (slow)~8.3GB4kllama.cpp › EvalPlus / EvalPerf published evaluation runsOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboard+2
665Yi-1.5 6B (base) · 4-bit 01.AI37±3local (slow)~5.4GB4kllama.cpp › BigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4Open Portuguese LLM Leaderboard (archived)EuroEval leaderboards (results collection), English generative leaderboard
666Qwen2-VL 2B Instruct · 4-bit Alibaba37±1local (slow)~3.3GB33kllama.cpp › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
667DeepSeek R1 Distill Qwen 1.5B · 4-bit DeepSeek37±2local (slow)~2.7GB131kllama.cpp › deepseek-ai/DeepSeek-R1 model cardCodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)BRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)Epoch AI — internal evaluationsTHINK-Bench: Evaluating Thinking Efficiency and Chain-of-Thought Quality of Large Reasoning Models (arXiv:2505.22113v1)EuroEval leaderboards (results collection), English generative leaderboard+1
668Gemma 2 2B IT · 4-bit Google37±1local (slow)~3.3GB8kllama.cpp › IBM Granite 3.0 Language Models technical report (GitHub)WildEval ZeroEval published resultsWildBench v2 leaderboard data (Ai2)LMArena leaderboard datasetSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)Open Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)+4
669Qwen2.5 Coder 0.5B Instruct · 4-bit Alibaba37±1local (slow)~1.9GB33kOllama › Qwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)Aider code editing benchmark runs (old edit leaderboard)
671Gemma 1.1 7B IT · 4-bit Google37±1local (slow)~7.1GB8kllama.cpp › EvalPlus / EvalPerf published evaluation runsLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard dataset+2
672Code Llama 7B Python · 4-bit Meta37±1local (slow)~5.8GB16kllama.cpp › Top Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsCode Llama: Open Foundation Models for Code — authors' own runsCRUXEval published evaluation results (authors' generations and scores)
673Dolphin 2.2.1 Mistral 7B · 4-bit Cognitive Computations37±2local (slow)~6.1GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard dataset
675Qwen1.5 7B Chat · 4-bit Alibaba37±1local (slow)~6.5GB32kllama.cpp › Top Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Open Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)MMLU-Pro author-released evaluation runsWildBench v2 leaderboard data (Ai2)+4
676Yi 1.5 6B Chat · 4-bit 01.AI37±2local (slow)~5.4GB4kllama.cpp › BigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4WildBench v2 leaderboard data (Ai2)SpeakLeash Polish EQ-Bench leaderboardOpen Portuguese LLM Leaderboard (archived)Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
677Falcon3 3B Instruct · 4-bit TII37±3local (slow)~3.6GB32kllama.cpp › Berkeley Function Calling Leaderboard result archiveCodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)Open Portuguese LLM Leaderboard (archived)
678Zephyr 7B alpha · 4-bit Hugging Face36±1local (slow)~6.1GBllama.cpp › LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Hallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetOpen Portuguese LLM Leaderboard (archived)
680Granite 3.1 2B Instruct · 4-bit IBM36±1local (slow)~3.1GB128kllama.cpp › ibm-granite/granite-3.3-8b-instruct model cardibm-granite/granite-3.1-3b-a800m-instruct model cardLMArena leaderboard datasetSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
681EXAONE 3.5 2.4B Instruct · 4-bit LG AI Research36±2local (slow)~3.2GB33kllama.cpp › LG AI Research vendor-reported benchmark scores for its own EXAONE models (Hugging Face model cards)SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)
683Code Llama 7B Instruct · 4-bit Meta36±1local (slow)~5.8GB16kllama.cpp › Top Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsCode Llama: Open Foundation Models for Code — authors' own runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsTulu 2 — base and instruction-tuned evaluationsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4
684LFM2 700M · 4-bit Liquid AI36±2local (slow)~2GB33kllama.cpp › LiquidAI/LFM2-1.2B model cardEuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
685Granite 3.2 2B Instruct · 4-bit IBM36±1local (slow)~3.1GB131kllama.cpp › ibm-granite/granite-3.3-8b-instruct model cardVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
686Mistral 7B Instruct v0.2 · 4-bit Mistral36±1local (slow)~6.1GB32kllama.cpp › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Hallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)+9
688Tulu 2 7B · 4-bit Allen Institute for AI36±2local (slow)~5.8GBllama.cpp › Tulu 2 — base and instruction-tuned evaluationsOpen Portuguese LLM Leaderboard (archived)
689xDAN-L1-Chat-RL-v1 · 4-bit xDAN-AI36±1local (slow)~6.1GBllama.cpp › EvalPlus / EvalPerf published evaluation runs
690Llama 3 8B (base) · 4-bit Meta36±1local (slow)~6.6GBllama.cpp › EvalPlus / EvalPerf published evaluation runsOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)BigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4Open Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)EuroEval leaderboards (results collection), English generative leaderboard
691Mistral 7B v0.1 (base) · 4-bit Mistral AI36±0local (slow)~6.1GB8kllama.cpp › EvalPlus / EvalPerf published evaluation runsStability AI — original StableLM zero-shot evaluation outputsMixtral of Experts — base-model evaluationsMiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies (Hu et al.)Mistral 7B — authors' own runsGemma report — original evaluations+4
692Qwen2.5 1.5B Instruct · 4-bit Alibaba36±1local (slow)~2.7GB33kllama.cpp › 2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsSmolLM2 paper (Hugging Face) — authors' own instruct-model evaluationsSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)SpeakLeash Polish EQ-Bench leaderboardOpen Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)+6
694Gemma 7B (base) · 4-bit Google35±1local (slow)~7.1GB8kllama.cpp › EvalPlus / EvalPerf published evaluation runsGemma report — original evaluationsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsCodeGemma: Open Code Models Based on Gemma — authors' own runsOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)+3
695Tulu 2 DPO 7B · 4-bit Allen Institute for AI35±1local (slow)~5.8GBllama.cpp › Tulu 2 — base and instruction-tuned evaluationsAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Open Portuguese LLM Leaderboard (archived)
696Eurus-2-7B-SFT · 4-bit PRIME-RL35±3†local (slow)~6.4GBllama.cpp › PRIME-RL Eurus-2-7B-PRIME model card, Evaluation table (own models)
698StarCoder2 7B · 4-bit BigCode35±1local (slow)~6.1GB16kllama.cpp › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsStarCoder 2 and The Stack v2: The Next Generation — authors' own runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4
699Granite 3.0 3B-A800M Instruct · 4-bit IBM35±1local (slow)~3.7GB4kllama.cpp › IBM Granite 3.0 Language Models technical report (GitHub)EuroEval leaderboards (results collection), English generative leaderboard
700DeepSeek LLM 7B Chat · 4-bit DeepSeek35±2†local (slow)~5.7GBOllama › DeepSeek vendor-reported benchmark scores for its own models (Hugging Face model cards)
702Granite 3.1 3B-A800M Instruct · 4-bit IBM35±2†local (slow)~3.6GB131kllama.cpp › ibm-granite/granite-3.1-3b-a800m-instruct model card
703Llama 2 13B (base) · 4-bit Meta34±1local (slow)~9.3GB4kOllama › Stability AI — original StableLM zero-shot evaluation outputsMixtral of Experts — base-model evaluationsMiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies (Hu et al.)Mistral 7B — authors' own runsTulu 2 — base and instruction-tuned evaluationsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)+3
704Code Llama 7B (base) · 4-bit Meta34±1local (slow)~5.8GB16kllama.cpp › EvalPlus / EvalPerf published evaluation runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsStarCoder 2 and The Stack v2: The Next Generation — authors' own runsMistral 7B — authors' own runsCode Llama: Open Foundation Models for Code — authors' own runsTulu 2 — base and instruction-tuned evaluations+2
706ChatGLM3 6B 128K · 4-bit Zhipu AI (THUDM)34±4local (slow)~5.8GB131kllama.cpp › NVIDIA RULER repository README results tableOpen Portuguese LLM Leaderboard (archived)
707Zephyr 7B beta · 4-bit Hugging Face34±1local (slow)~6.1GBllama.cpp › EvalPlus / EvalPerf published evaluation runsTulu 2 — base and instruction-tuned evaluationsLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Hallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)OpenChat 3.5 model card (openchat/openchat_3.5)+3
709Qwen3 0.6B · 4-bit Alibaba33±1local (slow)~2GB33kllama.cpp › Olmo 3LiquidAI/LFM2-1.2B model cardBerkeley Function Calling Leaderboard result archiveOpen Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)BRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)EuroEval leaderboards (results collection), English generative leaderboard+4
712StarCoder2 3B · 4-bit BigCode33±1local (slow)~3.4GB16kllama.cpp › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsStarCoder 2 and The Stack v2: The Next Generation — authors' own runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4
713Granite 3B Code Base (2K) · 4-bit IBM33±1local (slow)~3.7GBllama.cpp › Granite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4EuroEval leaderboards (results collection), English generative leaderboard
716Vicuna 7B v1.5 · 4-bit LMSYS33±1local (slow)~5.8GB4kllama.cpp › LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetLMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table+1
718Stable Code 3B · 4-bit Stability AI33±1local (slow)~3.3GB16kllama.cpp › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsStarCoder 2 and The Stack v2: The Next Generation — authors' own runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runs
719OLMo 2 1B Instruct (0425) · 4-bit Allen Institute for AI33±2local (slow)~2.4GB4kllama.cpp › 2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsOpen Portuguese LLM Leaderboard (archived)
720DeepSeek-Coder 1.3B Base · 4-bit DeepSeek32±1local (slow)~2.4GB16kOllama › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsStarCoder 2 and The Stack v2: The Next Generation — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4CRUXEval published evaluation results (authors' generations and scores)
722Gemma 7B IT · 4-bit Google32±1local (slow)~7.1GB8kllama.cpp › EvalPlus / EvalPerf published evaluation runsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)WildBench v2 leaderboard data (Ai2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)+5
723Mistral 7B Instruct v0.1 · 4-bit Mistral32±1local (slow)~6.1GB8kllama.cpp › Top Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)MMLU-Pro author-released evaluation runsPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetOpen Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)+2
724Qwen1.5 4B Chat · 4-bit Alibaba32±2local (slow)~4.1GB32kllama.cpp › LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetOpen Portuguese LLM Leaderboard (archived)EuroEval leaderboards (results collection), English generative leaderboard
725Ghost 7B Alpha · 4-bit Ghost X32±2local (slow)~6.1GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Open Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)
726CodeGemma 2B (v1.0) · 4-bit Google32±1local (slow)~3.2GBllama.cpp › EvalPlus / EvalPerf published evaluation runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsCodeGemma: Open Code Models Based on Gemma — authors' own runsBigCodeBench paper (Zhuo et al., ICLR 2025), arXiv 2406.15877v4
727Qwen-7B (base) · 4-bit Alibaba32±1local (slow)~6.6GB33kllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsOpen Portuguese LLM Leaderboard (archived)
728Phi-1.5 (1.3B) · 4-bit Microsoft32±1local (slow)~2.4GB2kllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsTextbooks Are All You Need II: phi-1.5 technical report — authors' own runsOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)Open Portuguese LLM Leaderboard (archived)
729Falcon-H1 1.5B Instruct · 4-bit TII31±4local (slow)~2.4GB131kllama.cpp › Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)Open Portuguese LLM Leaderboard (archived)
731Qwen2 1.5B Instruct · 4-bit Alibaba31±1local (slow)~2.6GB32kllama.cpp › WildEval ZeroEval published resultsQwen (Alibaba) vendor-reported benchmark scores for its own models (qwen.ai launch posts, Qwen2.5-VL and Qwen3-VL technical reports, Qwen3.8-Flash-Next card)Qwen2-0.5B-Instruct model cardLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsSpeakLeash Polish EQ-Bench leaderboardOpen Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)+1
732Qwen-7B-Chat · 4-bit Alibaba31±1local (slow)~6.6GB33kllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsOpen Portuguese LLM Leaderboard (archived)
733Gemma 3 1B · 4-bit Google31±1local (slow)~2.3GB32kllama.cpp › Google DeepMind Gemma 3 270M IT model card2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsLiquidAI/LFM2-1.2B model cardLiquidAI/LFM2.5-1.2B-Thinking model cardMIMIC Clinical Decision Making leaderboardGemma Team — Gemma 3 Technical Report+7
734StarCoderBase 7B · 4-bit BigCode31±1local (slow)~6GBOllama › EvalPlus / EvalPerf published evaluation runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsStarCoder 2 and The Stack v2: The Next Generation — authors' own runsCRUXEval published evaluation results (authors' generations and scores)
735Llama 2 7B (base) · 4-bit Meta31±1local (slow)~5.5GB4kOllama › Stability AI — original StableLM zero-shot evaluation outputsMixtral of Experts — base-model evaluationsOLMo: Accelerating the Science of Language Models — authors' own runsTextbooks Are All You Need II: phi-1.5 technical report — authors' own runsMistral 7B — authors' own runsMiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies (Hu et al.)+5
737LFM2 350M · 4-bit Liquid AI30±2local (slow)~1.7GB33kllama.cpp › LiquidAI/LFM2-1.2B model cardEuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
738SmolLM2 1.7B Instruct · 4-bit Hugging Face30±1local (slow)~2.7GB8kllama.cpp › 2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsSmolLM2 paper (Hugging Face) — authors' own instruct-model evaluationsLMArena leaderboard datasetSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)EuroEval leaderboards (results collection), English generative leaderboard
739Llama 3.2 1B Instruct · 4-bit Meta30±1local (slow)~2.3GB131kllama.cpp › IBM Granite 3.0 Language Models technical report (GitHub)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsLiquidAI/LFM2-1.2B model cardSmolLM2 paper (Hugging Face) — authors' own instruct-model evaluationsLiquidAI/LFM2.5-1.2B-Thinking model cardLMArena leaderboard dataset+13
740Falcon 7B (base) · 4-bit TII30±1local (slow)~5.9GB2kOllama › Stability AI — original StableLM zero-shot evaluation outputsOLMo: Accelerating the Science of Language Models — authors' own runsTextbooks Are All You Need II: phi-1.5 technical report — authors' own runsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)Towards a Holistic Evaluation of LLMs on Factual Knowledge Recall (arXiv:2404.16164v1)Open Portuguese LLM Leaderboard (archived)+1
741Llama 2 7B Chat · 4-bit Meta30±1local (slow)~5.8GB4kllama.cpp › Tulu 2 — base and instruction-tuned evaluationsLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Hallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)WildBench v2 leaderboard data (Ai2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)+11
746StableLM-3B-4E1T · 4-bit Stability AI29±1local (slow)~3.3GB4kllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsOpen Portuguese LLM Leaderboard (archived)
747Gemma 2B (base) · 4-bit Google29±1local (slow)~3.2GB8kllama.cpp › EvalPlus / EvalPerf published evaluation runsGemma report — original evaluationsTop Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM — authors' own runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsCodeGemma: Open Code Models Based on Gemma — authors' own runsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)+3
748OpenLLaMA 7B v2 · 8-bit OpenLM Research29±1local (slow)~9.1GBllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsOpen Portuguese LLM Leaderboard (archived)
749Gemma 1.1 2B IT · 4-bit Google29±1local (slow)~3.2GB8kllama.cpp › EvalPlus / EvalPerf published evaluation runsPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetSpeakLeash Polish EQ-Bench leaderboardArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Open Portuguese LLM Leaderboard (archived)+1
751Granite 3.1 1B-A400M Instruct · 4-bit IBM29±2†local (slow)~2.3GB131kllama.cpp › ibm-granite/granite-3.1-3b-a800m-instruct model card
753NexusRaven V2 13B · 4-bit Nexusflow29±3local (slow)~9.3GB16kOllama › Hallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)Berkeley Function Calling Leaderboard result archive
759StarCoderBase 3B · 4-bit BigCode28±1local (slow)~3.4GBOllama › EvalPlus / EvalPerf published evaluation runsStarCoder 2 and The Stack v2: The Next Generation — authors' own runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runs
760Granite 3.0 1B-A400M Instruct · 4-bit IBM27±1local (slow)~2.3GB4kllama.cpp › IBM Granite 3.0 Language Models technical report (GitHub)EuroEval leaderboards (results collection), English generative leaderboard
762ChatGLM3-6B · 4-bit Zhipu AI (THUDM)27±2local (slow)~5.8GBllama.cpp › Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools (Team GLM, arXiv 2406.12793v2)Open Portuguese LLM Leaderboard (archived)
763Granite 4.0 H 350M · 4-bit IBM27±1†local (slow)~1.7GB33kllama.cpp › ibm-granite/granite-4.0-h-1b model card
764Flan-T5 Large · 4-bit Google27±2local (slow)~4.8GBhow to run › Scaling Instruction-Finetuned Language Models — authors' own runsVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
766Gemma 2B IT · 4-bit Google27±1local (slow)~3.1GB8kllama.cpp › EvalPlus / EvalPerf published evaluation runsGranite Code Models: A Family of Open Foundation Models for Code Intelligence — authors' own runsWildBench v2 leaderboard data (Ai2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)+3
767EuroLLM 1.7B Instruct · 4-bit UTTER (EuroLLM)27±3local (slow)~2.6GB4kllama.cpp › SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)Open Portuguese LLM Leaderboard (archived)EuroEval leaderboards (results collection), English generative leaderboard
768Qwen2.5 0.5B Instruct · 4-bit Alibaba27±2local (slow)~2GB33kllama.cpp › HuggingFaceTB/SmolLM2-360M-Instruct model cardSpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)Open Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)Berkeley Function Calling Leaderboard result archiveVectara hallucination leaderboard (vectara/hallucination-leaderboard repository)+1
772Granite 4.0 350M · 4-bit IBM26±1local (slow)~1.7GB32kllama.cpp › ibm-granite/granite-4.0-h-1b model cardBerkeley Function Calling Leaderboard result archiveEpoch AI — internal evaluations
775Falcon 7B Instruct · 4-bit TII25±2local (slow)~5.9GBOllama › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Hallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)
776Qwen2 0.5B Instruct · 4-bit Alibaba25±2local (slow)~1.9GB32kllama.cpp › Qwen2-0.5B-Instruct model cardLiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsOpen Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)Open Portuguese LLM Leaderboard (archived)
777Pythia 12B · 4-bit EleutherAI25±1local (slow)~9.5GB2kllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsPythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsTowards a Holistic Evaluation of LLMs on Factual Knowledge Recall (arXiv:2404.16164v1)Open Portuguese LLM Leaderboard (archived)
778RedPajama-INCITE Base 3B v1 · 4-bit Together25±3local (slow)~3.4GB2kllama.cpp › RedPajama — original zero-shot 3B comparisonOpen Portuguese LLM Leaderboard (archived)
781OpenLLaMA 3B v2 · 4-bit OpenLM Research25±1local (slow)~4.2GBllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsOpen Portuguese LLM Leaderboard (archived)
784Falcon3 1B Instruct · 4-bit TII24±4local (slow)~2.7GB8kllama.cpp › Berkeley Function Calling Leaderboard result archiveCodeReviewQA: The Code Review Comprehension Assessment for Large Language Models (arXiv:2503.16167v2)Open Portuguese LLM Leaderboard (archived)
786Qwen1.5 1.8B Chat · 4-bit Alibaba24±2local (slow)~2.7GB33kOllama › LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)EuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
787Pythia 6.9B · 4-bit EleutherAI23±1local (slow)~6.1GB2kllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsOLMo: Accelerating the Science of Language Models — authors' own runsPythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsTowards a Holistic Evaluation of LLMs on Factual Knowledge Recall (arXiv:2404.16164v1)Open Portuguese LLM Leaderboard (archived)
788OpenAssistant SFT-4 Pythia 12B · 4-bit LAION (OpenAssistant)22±1local (slow)~9.5GBllama.cpp › InstructEval (DeCLaRe Lab) README results tableAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetLMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard tableLLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusion (arXiv:2306.02561v3)
789Pythia 2.8B · 4-bit EleutherAI21±1local (slow)~3.4GB2kllama.cpp › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsRedPajama — original zero-shot 3B comparisonOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)Open Portuguese LLM Leaderboard (archived)
790BLOOM 7.1B · 4-bit BigScience21±1local (slow)~7.1GB2kllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsBLOOM: A 176B-Parameter Open-Access Multilingual Language Model — authors' own runsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)Towards a Holistic Evaluation of LLMs on Factual Knowledge Recall (arXiv:2404.16164v1)Open Portuguese LLM Leaderboard (archived)From Test-Taking to Test-Making: Examining LLM Authoring of Commonsense Assessment Items (arXiv:2410.14897v1)
791OLMo 1B · 4-bit Allen Institute for AI21±2†local (slow)~2.2GB2kllama.cpp › OLMo: Accelerating the Science of Language Models — authors' own runs
792Phi-1 (1.3B) · 4-bit Microsoft21±4local (slow)~2.4GB2kllama.cpp › CRUXEval published evaluation results (authors' generations and scores)Open Portuguese LLM Leaderboard (archived)
793Qwen1.5 0.5B Chat · 4-bit Alibaba21±3local (slow)~1.9GB33kOllama › LiveBench model judgments (HF livebench/model_judgment, 2024-10-22 revision) with the LiveBench question datasetsEuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
795Pythia 2.8B (deduped) · 4-bit EleutherAI20±1local (slow)~3.4GB2kllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)Open Portuguese LLM Leaderboard (archived)
796Flan-T5 Base · 4-bit Google20±3†local (slow)~2.5GBhow to run › Scaling Instruction-Finetuned Language Models — authors' own runs
797Falcon-RW 1B · 16-bit TII20±2local (slow)~4.3GB2khow to run › Textbooks Are All You Need II: phi-1.5 technical report — authors' own runsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)
798OPT-2.7B · 4-bit Meta19±1local (slow)~7.1GB2khow to run › Stability AI — original StableLM zero-shot evaluation outputsPythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsOpen Portuguese LLM Leaderboard (archived)
799SmolLM2 360M Instruct · 8-bit Hugging Face19±3local (slow)~1.9GB8kllama.cpp › HuggingFaceTB/SmolLM2-360M-Instruct model cardEuroEval leaderboards (results collection), English generative leaderboard
802Pythia 1.4B · 4-bit EleutherAI18±2local (slow)~2.4GB2kllama.cpp › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)
803BLOOM 3B · 4-bit BigScience18±1local (slow)~4GB2kllama.cpp › Stability AI — original StableLM zero-shot evaluation outputsPythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsBLOOM: A 176B-Parameter Open-Access Multilingual Language Model — authors' own runsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)Open Portuguese LLM Leaderboard (archived)
806OPT-1.3B · 4-bit Meta16±1local (slow)~4.3GB2khow to run › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsTextbooks Are All You Need II: phi-1.5 technical report — authors' own runsOpen Portuguese LLM Leaderboard (archived)
807Gemma 3 270M IT · 8-bit Google15±4local (slow)~1.8GB32kllama.cpp › Google DeepMind Gemma 3 270M IT model cardEuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
808Pythia 1B · 4-bit EleutherAI14±1local (slow)~2.2GB2kllama.cpp › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsOLMo: Accelerating the Science of Language Models — authors' own runsOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)Open Portuguese LLM Leaderboard (archived)
809StableLM-Tuned-Alpha 7B · 4-bit Stability AI14±2local (slow)~6.9GBllama.cpp › Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMArena leaderboard datasetLMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard tableLLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusion (arXiv:2306.02561v3)
810BLOOM 1.7B · 4-bit BigScience14±1local (slow)~3.1GB2kllama.cpp › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsBLOOM: A 176B-Parameter Open-Access Multilingual Language Model — authors' own runsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)Open Portuguese LLM Leaderboard (archived)
811Flan-T5 Small · 4-bit Google14±3†local (slow)~1.8GBhow to run › Scaling Instruction-Finetuned Language Models — authors' own runs
812GPT-Neo 1.3B · 4-bit EleutherAI14±2local (slow)~7.1GB2khow to run › EleutherAI GPT-J-6B model card — EleutherAI lm-evaluation-harness runsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)EleutherAI GPT-Neo 2.7B model card — EleutherAI evaluation-harness runsOpen Portuguese LLM Leaderboard (archived)
813GPT-2 XL (1.5B) · 4-bit OpenAI11±1local (slow)~2.8GB1kllama.cpp › Surface Form Competition (Holtzman et al., EMNLP 2021) — authors' own GPT-2 runsTextbooks Are All You Need II: phi-1.5 technical report — authors' own runsEleutherAI GPT-J-6B model card — EleutherAI lm-evaluation-harness runsPROST (Aroca-Ouellette et al., Findings of ACL-IJCNLP 2021) - authors' own zero-shot runsEleutherAI GPT-Neo 2.7B model card — EleutherAI evaluation-harness runsOpen Medical-LLM Leaderboard results (HF dataset openlifescienceai/results)+2
814BLOOM 1.1B · 4-bit BigScience11±1local (slow)~2.6GB2kllama.cpp › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsBLOOM: A 176B-Parameter Open-Access Multilingual Language Model — authors' own runsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)
815Pythia 410M · 4-bit EleutherAI10±2local (slow)~1.8GB2kllama.cpp › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsOpen Portuguese LLM Leaderboard (archived)
818Zyte-1B · 4-bit venkycs (aihub-app)8±2local (slow)~2.2GBllama.cpp › EvalPlus / EvalPerf published evaluation runs
819GPT-Neo 125M · 4-bit EleutherAI5±5local (slow)~2.1GB2khow to run › Hallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)COPEN — EMNLP 2022 original zero-shot evaluationsOpen Portuguese LLM Leaderboard (archived)
820BLOOM 560M · 4-bit BigScience5±2local (slow)~2.1GB2kllama.cpp › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsBLOOM: A 176B-Parameter Open-Access Multilingual Language Model — authors' own runsHallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)Open Portuguese LLM Leaderboard (archived)
821OPT-350M · 4-bit Meta5±2local (slow)~2.2GB2khow to run › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsOpen Portuguese LLM Leaderboard (archived)
823Pythia 160M · 4-bit EleutherAI3±2local (slow)~1.6GB2kllama.cpp › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsOpen Portuguese LLM Leaderboard (archived)
824OPT-125M · 4-bit Meta3±2local (slow)~1.8GB2khow to run › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsOpen Portuguese LLM Leaderboard (archived)
825GPT-2 Large (774M) · 4-bit OpenAI1±2local (slow)~2GB1kllama.cpp › Surface Form Competition (Holtzman et al., EMNLP 2021) — authors' own GPT-2 runsPROST (Aroca-Ouellette et al., Findings of ACL-IJCNLP 2021) - authors' own zero-shot runsCOPEN — EMNLP 2022 original zero-shot evaluationsEuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
826GPT-2 Medium (355M) · 4-bit OpenAI-1±2local (slow)~1.8GB1kllama.cpp › Surface Form Competition (Holtzman et al., EMNLP 2021) — authors' own GPT-2 runsPROST (Aroca-Ouellette et al., Findings of ACL-IJCNLP 2021) - authors' own zero-shot runsCOPEN — EMNLP 2022 original zero-shot evaluationsEuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
827Pythia 70M · 4-bit EleutherAI-7±2local (slow)~1.6GB2kllama.cpp › Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling — authors' own runsOpen Portuguese LLM Leaderboard (archived)
828GPT-2 Small (124M) · 4-bit OpenAI-9±3local (slow)~1.6GB1kllama.cpp › Surface Form Competition (Holtzman et al., EMNLP 2021) — authors' own GPT-2 runsPROST (Aroca-Ouellette et al., Findings of ACL-IJCNLP 2021) - authors' own zero-shot runsCOPEN — EMNLP 2022 original zero-shot evaluationsEuroEval leaderboards (results collection), English generative leaderboardOpen Portuguese LLM Leaderboard (archived)
—K2 Horizon 7B · gguf-q4_k_m Institute of Foundation Models (MBZUAI)not scored yetlocal (slow)~7.7GB524kllama.cpp ›

+ 535 more don't fit your machine or budget · show them

+ 112 models with fewer than 3 results left off, since one or two tests say little about overall ability · show them

830 of 1134 catalogue models listed here · sources and licences

* new or thinly tested: ranked on too few benchmark results · † only the maker’s own numbers so far · ~RAM is estimated; full context may need more memory