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

Best AI models for chat and writing

A leaderboard of the best AI models for chat and writing. Gemini 4 Argon (limited preview) is #1 right now, followed by Claude Fable 5 and Claude Opus 4.6. Models are ranked on human votes from LMArena and on writing and instruction-following tests. Filter by your computer or budget to find the best one you can use.

471 models · scores combine 19 benchmarks · updated how it works · find a leaderboard · benchmark your machine

Scores use reviewed chat & writing evaluations only. Compare scores within this task. Models without task measurements are excluded.

Filter by your hardware and budget · reset
+ more
$ terminal

detecting your OS…

Score vs API price

#modeltask score↓where↓speed↓RAM↓context↓get itsources
1Gemini 4 Argon (limited preview) Google78±1*limited previewLMArena leaderboard dataset
2Claude Fable 5 Anthropic76±1*api $20.00/M1Mdocs.anthropic.com ↗LMArena leaderboard dataset
3Claude Opus 4.6 Anthropic76±1*api $10.00/M1Mdocs.anthropic.com ↗LMArena leaderboard datasetmeituan-longcat/LongCat-2.0 model card
4Claude Opus 5.5 Anthropic76±1*api $8.00/M1Mdocs.anthropic.com ↗LMArena leaderboard dataset
5Claude Fable 5.1 Anthropic75±1*api $20.00/M1Mdocs.anthropic.com ↗LMArena leaderboard dataset
6Claude Opus 4.7 Anthropic75±1*api $10.00/M1Mdocs.anthropic.com ↗LMArena leaderboard datasetmeituan-longcat/LongCat-2.0 model card
7Gemini 3.8 Flash Google75±1*api $1.50/M1Mai.google.dev ↗LMArena leaderboard dataset
8Muse Spark 1.3 Meta75±1*api $2.00/M1Mdev.meta.ai ↗LMArena leaderboard dataset
9Muse Spark 1.2 Meta74±1*api $2.00/M1Mdev.meta.ai ↗LMArena leaderboard dataset
10Muse Spark 1.1 Meta74±1*api $2.00/M1Mdev.meta.ai ↗LMArena leaderboard dataset
11Claude Opus 5 Anthropic74±1*api $10.00/M1Mdocs.anthropic.com ↗LMArena leaderboard dataset
12Muse Spark Meta74±1*comparison onlyLMArena leaderboard dataset
13Kimi K3 Moonshot AI74±1*api $6.00/M +local~884GB1Mplatform.moonshot.ai ↗LMArena leaderboard dataset
14Gemini 3.7 Flash Google74±1*api $1.50/M1Mai.google.dev ↗LMArena leaderboard dataset
15Gemini 3.1 Pro (Preview) Google74±1*api $4.50/M1Mai.google.dev ↗LMArena leaderboard datasetmeituan-longcat/LongCat-2.0 model card
16Gemini 3 Pro (Preview) Google73±1*comparison only1MLMArena leaderboard dataset
17GPT-5.6 Sol OpenAI73±1*api $8.00/M1.1Mplatform.openai.com ↗LMArena leaderboard dataset
18Gemini 3.6 Flash Google73±1*api $1.50/M1Mai.google.dev ↗LMArena leaderboard dataset
19GPT-6.1 Sol OpenAI73±1*api $4.00/M1.1Mdevelopers.openai.com ↗LMArena leaderboard dataset
20Qwen3.8 Max (2.4T-A95B) Alibaba73±1*api $3.00/M +local~380GB262kdeepinfra.com ↗LMArena leaderboard dataset
21GPT-5.5 OpenAI73±1*api $11.25/M1.1Mplatform.openai.com ↗LMArena leaderboard datasetmeituan-longcat/LongCat-2.0 model card
22MiMo-V2.6-Pro Xiaomi73±1*api $0.54/M +local~443GB1Mmimo.mi.com ↗LMArena leaderboard dataset
23Claude Opus 4.8 Anthropic73±1*api $10.00/M1Mdocs.anthropic.com ↗LMArena leaderboard datasetmeituan-longcat/LongCat-2.0 model card
24GLM-5.3 Z.ai73±1*api $2.15/M +local~239GB1Mdocs.z.ai ↗LMArena leaderboard dataset
25GPT-6 Astra OpenAI72±1*api $20.00/M1.1Mplatform.openai.com ↗LMArena leaderboard dataset
26Gemini 3.5 Flash Google72±1*api $3.38/M1Mai.google.dev ↗LMArena leaderboard dataset
27GLM-5.2 Z.ai72±1*api $2.15/M +local~121GB1Mdocs.z.ai ↗LMArena leaderboard dataset
28GPT-5.4 OpenAI72±1*api $5.63/M1.1Mplatform.openai.com ↗LMArena leaderboard dataset
29DeepSeek V4.1 Flash DeepSeek72±1*api $0.26/M +local~537GB1Mapi-docs.deepseek.com ↗LMArena leaderboard dataset
30GLM-5.3-Flash Z.ai72±1*api $0.24/M +local~103GB1Mdocs.z.ai ↗LMArena leaderboard dataset
31Claude Opus 4.5 Anthropic72±1*api $10.00/M200kdocs.anthropic.com ↗LMArena leaderboard dataset
32GPT-5.5 Instant OpenAI72±1*comparison onlyLMArena leaderboard dataset
33Claude Sonnet 4.6 Anthropic72±1*api $6.00/M1Mdocs.anthropic.com ↗LMArena leaderboard dataset
34Gemini 3 Flash (Preview) Google72±1*api $1.13/M1Mai.google.dev ↗LMArena leaderboard dataset
35Grok 4.20 (0309 reasoning) SpaceXAI (xAI)72±1*api $1.56/M1Mdocs.x.ai ↗LMArena leaderboard dataset
36Grok 4.20 Multi-Agent (0309) SpaceXAI (xAI)72±1*api $1.56/M1Mdocs.x.ai ↗LMArena leaderboard dataset
37Claude Sonnet 5.5 Anthropic72±1*api $4.00/M1Mplatform.claude.com ↗LMArena leaderboard dataset
38MiMo-V2.5-Pro (1T-A42B) Xiaomi71±1*api $0.65/M +local~163GB1Mnovita.ai ↗LMArena leaderboard dataset
39GPT-5.6 Terra OpenAI71±1*api $4.50/M1.1Mplatform.openai.com ↗LMArena leaderboard dataset
40Grok 4.5 SpaceXAI (xAI)71±1*api $3.00/M500kdocs.x.ai ↗LMArena leaderboard dataset
41DeepSeek V4 Pro (0813) DeepSeek71±1*api $1.63/M +local~1010GB1Mdeepinfra.com ↗LMArena leaderboard dataset
42GLM-5.1 Z.ai71±1*api $2.15/M +local~119GB200kdocs.z.ai ↗LMArena leaderboard datasetLGAI-EXAONE/K-EXAONE-2.0-750B-A37B model card
43Claude Sonnet 5 Anthropic70±1*api $4.00/M1Mdocs.anthropic.com ↗LMArena leaderboard dataset
44Kimi K2.6 Moonshot AI70±1*api $1.71/M +local~317GB262kplatform.moonshot.ai ↗LMArena leaderboard dataset
45Qwen3.6 Max Preview Alibaba70±1*api $2.92/M256kwww.alibabacloud.com ↗LMArena leaderboard dataset
46GPT-6 Sol OpenAI70±1*api $4.00/M1.1Mplatform.openai.com ↗LMArena leaderboard dataset
47GLM-5 Z.ai70±1*api $1.55/M +local~119GB205kdocs.z.ai ↗LMArena leaderboard dataset
48Claude Sonnet 4.5 Anthropic70±1*api $6.00/M200kdocs.anthropic.com ↗LMArena leaderboard dataset
49Dola Seed 2.0 Pro ByteDance70±1*comparison onlyLMArena leaderboard dataset
50GPT-5.1 OpenAI70±1*api $3.44/M400kplatform.openai.com ↗LMArena leaderboard dataset
51Qwen3.7 Plus Alibaba70±1*api $0.70/M1Mwww.alibabacloud.com ↗LMArena leaderboard dataset
52Hunyuan Hy3 (295B-A21B) Tencent70±1*api $0.23/M +local~96GB256kdeepinfra.com ↗LMArena leaderboard dataset
53Gemini 3.5 Flash-Lite Google70±1*api $0.85/M1Mai.google.dev ↗LMArena leaderboard dataset
54DeepSeek V4 Pro (Preview, Apr 2026) DeepSeek69±1*local~604GBllama.cpp › LMArena leaderboard datasetLGAI-EXAONE/K-EXAONE-2.0-750B-A37B model card
55Grok 4.6 SpaceXAI (xAI)69±1*api $3.00/M500kdocs.x.ai ↗LMArena leaderboard dataset
56Step 5 Preview StepFun69±1*comparison onlyLMArena leaderboard dataset
57Gemma 4 31B Google69±1*api $0.19/M +local~21GB262kdeepinfra.com ↗LMArena leaderboard datasetGemma Team — Gemma 4 Technical Report
58GPT-5.6 Luna OpenAI69±1*api $0.45/M1.1Mplatform.openai.com ↗LMArena leaderboard dataset
59MiMo-V2.6-Flash (309B-A15B) Xiaomi69±1*api $0.17/M +local~136GB1Mmimo.mi.com ↗LMArena leaderboard dataset
60Kimi K2.5 Moonshot AI69±1*api $1.20/M +local~160GB262kdocs.aws.amazon.com ↗LMArena leaderboard dataset
61Claude Opus 4.1 Anthropic69±1*comparison onlyLMArena leaderboard dataset
62MiMo-V2-Pro Xiaomi69±1*comparison onlyLMArena leaderboard dataset
63GPT-5.4 mini OpenAI69±1*api $1.69/M400kplatform.openai.com ↗LMArena leaderboard dataset
64Gemini 2.5 Pro Google68±1*api $3.44/M1Mai.google.dev ↗LMArena leaderboard datasetMeituan LongCat-Flash-Thinking model card (evaluation table)
65Qwen3.6 Plus Alibaba68±1*api $1.13/M1Mdocs.together.ai ↗LMArena leaderboard dataset
66GPT-6 Luna OpenAI68±1*api $0.20/M1.1Mplatform.openai.com ↗LMArena leaderboard dataset
67Grok 4.7 SpaceXAI (xAI)68±1*api $3.00/M500kdocs.x.ai ↗LMArena leaderboard dataset
68ChatGPT-4o (2025-03-26) OpenAI68±1*comparison onlyLMArena leaderboard dataset
69Grok 4.3 SpaceXAI (xAI)68±1*api $1.56/M1Mdocs.x.ai ↗LMArena leaderboard dataset
70GLM-4.7 Z.ai68±1*api $1.00/M +local~58GB205kdocs.z.ai ↗LMArena leaderboard dataset
71Inkling (975B-A41B) Thinking Machines68±1*api $1.73/M +local~156GB1Mdeepinfra.com ↗LMArena leaderboard dataset
72Nemotron Cascade 2 30B-A3B NVIDIA68±4*†local~20GB256kLM Studio › NVIDIA — Nemotron-Cascade 2 technical report
73Qwen3.5 397B-A17B Alibaba68±1*api $1.09/M +local~65GB262kdeepinfra.com ↗LMArena leaderboard datasetQwen/Qwen3.5-397B-A17B model card
74MiniMax-M3 MiniMax68±1*api $0.48/M +local~137GB1Mdeepinfra.com ↗LMArena leaderboard dataset
75GPT-4.5 Preview OpenAI68±1comparison onlyLMArena leaderboard datasetArena-Hard-Auto v2.0-Preview leaderboard (repository README)
76DeepSeek V4 Flash (Preview, Apr 2026) DeepSeek67±1*local~103GBllama.cpp › LMArena leaderboard dataset
77Qwen3.8 27B Alibaba67±1*api $1.60/M +local~18GB262kconsole.groq.com ↗LMArena leaderboard dataset
78Gemma 4 26B-A4B Google67±1*api $0.14/M +local~18GB262kdeepinfra.com ↗LMArena leaderboard datasetGemma Team — Gemma 4 Technical Report
79GPT-5.2 OpenAI67±1*api $4.81/M400kplatform.openai.com ↗LMArena leaderboard dataset
80o3 OpenAI67±1api $3.50/M200kplatform.openai.com ↗LMArena leaderboard datasetQwen/Qwen3-235B-A22B-Thinking-2507 model cardArena-Hard-Auto v2.0-Preview leaderboard (repository README)Meituan LongCat-Flash-Thinking model card (evaluation table)
81GPT-5 OpenAI67±1*api $3.44/M400kplatform.openai.com ↗LMArena leaderboard dataset
82MiMo-V2.5 (310B-A15B) Xiaomi67±1*api $0.21/M +local~100GB1Mnovita.ai ↗LMArena leaderboard dataset
83Gemma 4 12B Google67±6*local~8.5GB262kLM Studio › XHToken/Spark-X2.5-4B model cardGemma Team — Gemma 4 Technical Report
84GLM-5V-Turbo Z.ai67±1*api $1.90/M200kwww.siliconflow.com ↗LMArena leaderboard dataset
85Gemini 3.1 Flash-Lite Google67±1*api $0.56/M1Mai.google.dev ↗LMArena leaderboard dataset
86MiMo-V2-Omni Xiaomi66±1*comparison only262kLMArena leaderboard dataset
87Kimi K2 Thinking Moonshot AI66±1*api $1.07/M +local~160GB262kdocs.aws.amazon.com ↗LMArena leaderboard dataset
88Grok 4.1 Fast (Reasoning) SpaceXAI (xAI)66±1*api $0.28/M128kazure.microsoft.com ↗LMArena leaderboard dataset
89Mistral Large 4 (public preview) Mistral66±1*api $2.06/M1Mdocs.mistral.ai ↗LMArena leaderboard dataset
90Mistral Medium 3.5 (128B) Mistral66±1*api $3.00/M +local~42GB262kdocs.mistral.ai ↗LMArena leaderboard dataset
91Nemotron 3 Ultra 550B-A55B NVIDIA66±1*api $1.35/M +local~89GB1Mdocs.together.ai ↗LMArena leaderboard dataset
92Claude Opus 4 Anthropic66±1*comparison onlyLMArena leaderboard dataset
93Muse Glimmer 30B Meta66±1*api $0.64/M +local~19GB131kfireworks.ai ↗LMArena leaderboard dataset
94DeepSeek V3.2 Exp DeepSeek66±1*local~726GBhow to run › LMArena leaderboard dataset
95GLM-4.6 Z.ai66±1*api $1.00/M +local~58GB205kdocs.z.ai ↗LMArena leaderboard dataset
96DeepSeek V3.2 DeepSeek65±1*api $0.29/M +local~110GB128kdeepinfra.com ↗LMArena leaderboard dataset
97Qwen3-Max (2025-09-23) Alibaba65±1*api $2.40/M262kwww.alibabacloud.com ↗LMArena leaderboard dataset
98DeepSeek R1 0528 DeepSeek65±1*api $0.91/M +local~110GB164kdeepinfra.com ↗LMArena leaderboard dataset
99Qwen3 235B-A22B Instruct 2507 Alibaba65±1api $0.20/M +local~76GB262kdeepinfra.com ↗LMArena leaderboard datasetQwen3-VL-235B-A22B-Instruct model cardQwen/Qwen3-Next-80B-A3B-Instruct model cardMeituan LongCat-Flash-Chat model card (evaluation table)
100Kimi K2 Instruct (0905) Moonshot AI65±1*local~298GBllama.cpp › LMArena leaderboard dataset
101Kimi K2 Instruct (0711) Moonshot AI65±1*api $1.00/M +local~160GB131knovita.ai ↗LMArena leaderboard datasetMeituan LongCat-Flash-Chat model card (evaluation table)
102DeepSeek V3.1 Terminus DeepSeek65±1*local~230GB164kllama.cpp › LMArena leaderboard dataset
103Qwen3.5 122B-A10B Alibaba65±1*api $1.10/M +local~75GB262kwww.alibabacloud.com ↗LMArena leaderboard datasetQwen/Qwen3.5-122B-A10B model card
104MiniMax-M2.7 MiniMax64±1*api $0.53/M +local~75GB205kplatform.minimax.io ↗LMArena leaderboard dataset
105Claude Haiku 4.5 Anthropic64±1*api $2.00/M200kdocs.anthropic.com ↗LMArena leaderboard dataset
106Qwen3-Max (2026-01-23) Alibaba64±5*†comparison onlyQwen/Qwen3.5-397B-A17B model card
107DeepSeek V3.1 DeepSeek64±1api $0.42/M +local~110GB131kdeepinfra.com ↗LMArena leaderboard datasetMeituan LongCat-Flash-Chat model card (evaluation table)Meituan LongCat-Flash-Thinking model card (evaluation table)
108Mistral Large 3 (675B-A41B) Mistral64±1*api $0.75/M +local~108GB262kdocs.mistral.ai ↗LMArena leaderboard datasetmistralai/Mistral-Small-4-119B-2603 model card
109Qwen3-VL 235B-A22B Instruct Alibaba64±1local~134GB256kLM Studio › LMArena leaderboard datasetQwen3-VL-235B-A22B-Instruct model card
110Grok 4 (0709) SpaceXAI (xAI)64±1*comparison onlyLMArena leaderboard dataset
111Spark-X2.5 4B iFLYTEK (SparkLLM)64±6*†local~3.9GB1MLM Studio › XHToken/Spark-X2.5-4B model card
112GPT-4.1 OpenAI64±1api $3.50/M1Mplatform.openai.com ↗LMArena leaderboard datasetArena-Hard-Auto v2.0-Preview leaderboard (repository README)Meituan LongCat-Flash-Chat model card (evaluation table)
113Qwen3.5 27B Alibaba64±1*api $0.84/M +local~18GB262kdeepinfra.com ↗LMArena leaderboard datasetQwen/Qwen3.5-122B-A10B model card
114GLM-4.5 Z.ai64±1api $1.00/M +local~58GB131kdocs.z.ai ↗LMArena leaderboard datasetMeituan LongCat-Flash-Thinking model card (evaluation table)GLM-4.5 Team (Zhipu AI & Tsinghua) — GLM-4.5 technical report
115Hy3 preview (295B-A21B) Tencent64±1*local~629GBhow to run › LMArena leaderboard dataset
116Mistral Medium 3.1 (2508) Mistral64±1*comparison onlyLMArena leaderboard dataset
117Gemini 2.5 Flash Google63±1*api $0.85/M1Mai.google.dev ↗LMArena leaderboard datasetMeituan LongCat-Flash-Chat model card (evaluation table)
118K-EXAONE 2.0 (750B-A37B) LG AI Research63±4*†local~791GB262khow to run › LGAI-EXAONE/K-EXAONE-2.0-750B-A37B model card
119Inkling Small (266B-A12B) Thinking Machines63±1*api $0.64/M +local~86GB1Mdeepinfra.com ↗LMArena leaderboard dataset
120Grok 4 Fast SpaceXAI (xAI)63±1*comparison onlyLMArena leaderboard dataset
121o1 (2024-12-17) OpenAI63±1api $26.25/M200kdevelopers.openai.com ↗LMArena leaderboard datasetArena-Hard-Auto v2.0-Preview leaderboard (repository README)
122Gemini 2.5 Flash Preview (09-2025) Google63±1*comparison onlyLMArena leaderboard dataset
123LongCat-Flash-Chat (560B-A27B) Meituan63±1*local~333GB128kLM Studio › LMArena leaderboard datasetMeituan LongCat-Flash-Chat model card (evaluation table)
124Nemotron-Cascade-8B NVIDIA63±7*†local~6.3GBLM Studio › NVIDIA — Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models
125GPT-5.4 nano OpenAI63±1*api $0.46/M400kplatform.openai.com ↗LMArena leaderboard dataset
126Qwen3 235B-A22B Thinking 2507 Alibaba63±1api $0.97/M +local~76GB131knovita.ai ↗LMArena leaderboard datasetQwen/Qwen3-Next-80B-A3B-Thinking model cardQwen/Qwen3-235B-A22B-Thinking-2507 model cardQwen3-VL-235B-A22B-Thinking model cardMeituan LongCat-Flash-Thinking model card (evaluation table)
127Claude Sonnet 4 Anthropic63±1*comparison onlyLMArena leaderboard datasetMeituan LongCat-Flash-Chat model card (evaluation table)
128Qwen3-Next 80B-A3B Instruct Alibaba63±1api $0.34/M +local~27GB131kdeepinfra.com ↗LMArena leaderboard datasetQwen/Qwen3-Next-80B-A3B-Instruct model card
129DeepSeek R1 (Jan '25) DeepSeek63±1api $4.00/M +local~110GB128kdocs.together.ai ↗LMArena leaderboard datasetdeepseek-ai/DeepSeek-R1 model cardArena-Hard-Auto v2.0-Preview leaderboard (repository README)OpenCompass CompassAcademic leaderboard (full version) Space: committed predictions file
130Qwen3.5 Flash Alibaba62±1*api $0.17/M1Mwww.alibabacloud.com ↗LMArena leaderboard dataset
131LFM2.5 2.6B Liquid AI62±7*local~3.2GB131kLM Studio › openbmb/MiniCPM5-2B model card
132o4-mini OpenAI62±1api $1.93/M200kplatform.openai.com ↗LMArena leaderboard datasetQwen/Qwen3-235B-A22B-Thinking-2507 model cardArena-Hard-Auto v2.0-Preview leaderboard (repository README)
133DeepSeek V3 0324 DeepSeek62±1*local~108GB164kllama.cpp › LMArena leaderboard datasetOpenCompass CompassAcademic leaderboard (full version) Space: committed predictions file
134Step 3.5 Flash (196B-A11B) StepFun62±1*api $0.15/M +local~118GB256kplatform.stepfun.ai ↗LMArena leaderboard dataset
135Qwen3-VL 235B-A22B Thinking Alibaba62±1*local~111GB256kLM Studio › LMArena leaderboard datasetQwen3-VL-235B-A22B-Thinking model card
136Qwen3.5 35B-A3B Alibaba62±1api $0.35/M +local~23GB262kdeepinfra.com ↗LMArena leaderboard datasetNVIDIA — Nemotron-Cascade 2 technical reportQwen/Qwen3.5-122B-A10B model card
137MiniMax-M2.5 MiniMax61±1*api $0.53/M +local~75GB205kplatform.minimax.io ↗LMArena leaderboard dataset
138Claude 3.7 Sonnet Anthropic61±1comparison onlyLMArena leaderboard datasetArena-Hard-Auto v2.0-Preview leaderboard (repository README)OpenCompass CompassAcademic leaderboard (full version) Space: committed predictions file
139GPT-5 mini OpenAI61±1*api $0.69/M400kplatform.openai.com ↗LMArena leaderboard dataset
140o1-preview OpenAI61±1*comparison onlyLMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
141Qwen3-Coder 480B-A35B Instruct Alibaba61±1*api $0.67/M +local~78GB262knovita.ai ↗LMArena leaderboard dataset
142Mistral Medium 3 (2505) Mistral61±1*comparison onlyLMArena leaderboard dataset
143Solar Pro 4 Upstage61±1*api $0.53/M524kdevelopers.upstage.ai ↗LMArena leaderboard dataset
144MiMo-V2-Flash (309B-A15B) Xiaomi61±1*local~108GBllama.cpp › LMArena leaderboard dataset
145MiniMax-M2.1 MiniMax61±1*local~79GBllama.cpp › LMArena leaderboard dataset
146LFM2.5 8B-A1B Liquid AI60±5*local~6.5GB128kLM Studio › LiquidAI/LFM2.5-8B-A1B model cardopenbmb/MiniCPM5-2B model card
147LongCat-Flash-Thinking (560B-A27B) Meituan60±5*†comparison only128kMeituan LongCat-Flash-Thinking model card (evaluation table)
148Qwen3 30B-A3B Instruct 2507 Alibaba60±1api $0.15/M +local~20GB262ktokenfactory.nebius.com ↗LMArena leaderboard datasetQwen/Qwen3-Next-80B-A3B-Instruct model cardQwen/Qwen3-30B-A3B-Instruct-2507 model card
149GPT-4.1 mini OpenAI60±1api $0.70/M1Mplatform.openai.com ↗LMArena leaderboard datasetArena-Hard-Auto v2.0-Preview leaderboard (repository README)
150GLM-4.6V Z.ai60±1*api $0.45/M +local~51GB128kdocs.z.ai ↗LMArena leaderboard dataset
151Trinity-Large-Preview Arcee AI60±1*local~155GBllama.cpp › LMArena leaderboard dataset
152LongCat 2.0 (1.6T-A48B) Meituan60±7*†api $1.30/M +local~2155GB1Mlongcat.chat ↗meituan-longcat/LongCat-2.0 model card
153K-EXAONE (236B-A23B) LG AI Research60±3*†local~152GB262kllama.cpp › LGAI-EXAONE/K-EXAONE-2.0-750B-A37B model cardLG AI Research vendor-reported benchmark scores for its own EXAONE models (Hugging Face model cards)
154Spark-X2.5 1.7B iFLYTEK (SparkLLM)60±5*†local~2.7GB1Mllama.cpp › XHToken/Spark-X2.5-4B model card
155Gemini 2.5 Flash-Lite Preview (06-17) Google59±1*comparison onlyLMArena leaderboard dataset
156Qwen3 235B-A22B (Apr 2025, hybrid) Alibaba59±1api $1.23/M +local~76GB131kwww.alibabacloud.com ↗LMArena leaderboard datasetQwen/Qwen3-235B-A22B-Thinking-2507 model cardQwen/Qwen3-30B-A3B-Thinking-2507 model cardQwen/Qwen3-30B-A3B-Instruct-2507 model cardArena-Hard-Auto v2.0-Preview leaderboard (repository README)
157GLM-4.5-Air Z.ai59±1*api $0.42/M +local~35GB131kdocs.z.ai ↗LMArena leaderboard datasetGLM-4.5 Team (Zhipu AI & Tsinghua) — GLM-4.5 technical report
158Qwen2.5-Max (2025-01-25) Alibaba59±1*comparison onlyLMArena leaderboard dataset
159Gemma 2 9B IT DPO (Princeton) Princeton NLP59±2*local~7.6GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
160Qwen3-Next 80B-A3B Thinking Alibaba59±1api $0.41/M +local~27GB131kwww.alibabacloud.com ↗LMArena leaderboard datasetQwen/Qwen3-Next-80B-A3B-Thinking model card
161o3-mini OpenAI59±1api $1.93/M200kdevelopers.openai.com ↗LMArena leaderboard datasetArena-Hard-Auto v2.0-Preview leaderboard (repository README)Microsoft — Phi-4-reasoning Technical Report
162Claude 3.5 Sonnet (2024-10-22) Anthropic59±1comparison onlyLMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Arena-Hard-Auto v2.0-Preview leaderboard (repository README)OpenCompass CompassAcademic leaderboard (full version) Space: committed predictions file
163EXAONE 4.5 33B LG AI Research59±4*†local~24GB262kllama.cpp › LG AI Research vendor-reported benchmark scores for its own EXAONE models (Hugging Face model cards)
164LFM2.5 1.2B Thinking Liquid AI59±6*†local~2.2GB33kLM Studio › LiquidAI/LFM2.5-1.2B-Thinking model card
165Trinity-Large-Thinking Arcee AI58±1*api $0.39/M +local~141GB512kdocs.arcee.ai ↗LMArena leaderboard dataset
166GLM-4.7-Flash Z.ai58±1*api $0.0000/M +local~19GB200kdocs.z.ai ↗LMArena leaderboard dataset
167MiniMax-M1 (80k) MiniMax58±1*api $0.96/M +local~959GB1Mnovita.ai ↗LMArena leaderboard dataset
168Grok 3 Mini Beta SpaceXAI (xAI)58±1*comparison onlyLMArena leaderboard dataset
169Qwen3.5 9B Alibaba58±3api $0.11/M +local~7.4GB262kdeepinfra.com ↗JetBrains/Mellum2-12B-A2.5B-Thinking model cardJetBrains/Mellum2-12B-A2.5B-Instruct model cardQwen/Qwen3.5-9B model cardJetBrains/Mellum2.1-12B-A2.5B-Thinking model card
170Gemma 3 27B Google58±1api $0.10/M +local~18GB131kdeepinfra.com ↗LMArena leaderboard datasetArena-Hard-Auto v2.0-Preview leaderboard (repository README)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsGemma Team — Gemma 4 Technical Report
171Qwen3 30B-A3B Thinking 2507 Alibaba58±2api $0.75/M +local~20GB262kwww.alibabacloud.com ↗Qwen/Qwen3-Next-80B-A3B-Thinking model cardQwen/Qwen3-30B-A3B-Thinking-2507 model cardByteDance-Seed/Seed-OSS-36B-Instruct model card
172Nemotron 3 Super 120B-A12B NVIDIA58±1*api $0.28/M +local~73GB262kdocs.aws.amazon.com ↗LMArena leaderboard dataset
173DeepSeek V3 (Dec '24) DeepSeek57±1api $0.46/M +local~337GB131kdeepinfra.com ↗LMArena leaderboard datasetdeepseek-ai/DeepSeek-R1 model cardOpenCompass CompassAcademic leaderboard (full version) Space: committed predictions file
174Gemini 2.0 Flash (001) Google57±1*comparison onlyLMArena leaderboard dataset
175Granite 4.1 30B IBM57±5*†local~19GBOllama › ibm-granite/granite-4.1-8b model card
176Phi-4-reasoning-plus 14B Microsoft57±3†local~10GBLM Studio › Microsoft — Phi-4-reasoning Technical Reportmicrosoft/Phi-4-reasoning model card
177Qwen3-VL 32B Thinking Alibaba57±3*†api $0.28/M +local~22GB256kwww.alibabacloud.com ↗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)
178Mistral Small 3.2 24B Mistral57±1*api $0.11/M +local~16GB128kdeepinfra.com ↗LMArena leaderboard datasetmistralai/Mistral-Small-3.2-24B-Instruct-2506 model card
179INTELLECT-3 Prime Intellect57±1*local~48GBllama.cpp › LMArena leaderboard dataset
180Command A (03-2025) Cohere57±1*api $4.38/M +local~37GB256kazure.microsoft.com ↗LMArena leaderboard dataset
181Gemini 2.0 Flash-Lite Preview (02-05) Google57±1*comparison onlyLMArena leaderboard dataset
182GLM-4.5V Z.ai57±1*api $0.90/M +local~69GB64kdocs.z.ai ↗LMArena leaderboard dataset
183gpt-oss-120b OpenAI56±1*api $0.26/M +local~67GB131kconsole.groq.com ↗LMArena leaderboard dataset
184Gemini 1.5 Pro (002) Google56±1*comparison onlyLMArena leaderboard dataset
185Llama 3.1 Nemotron Ultra 253B v1 NVIDIA56±1local~82GB128kllama.cpp › LMArena leaderboard datasetNVIDIA — Llama-Nemotron: Efficient Reasoning Models
186Nemotron 3.5 Lightning 30B-A3B NVIDIA56±1*api $0.087/M +local~20GB262kfireworks.ai ↗LMArena leaderboard dataset
187Qwen3.5 4B Alibaba56±3local~4.3GB262kllama.cpp › JetBrains/Mellum2-12B-A2.5B-Thinking model cardJetBrains/Mellum2-12B-A2.5B-Instruct model cardQwen/Qwen3.5-9B model cardopenbmb/MiniCPM5-2B model card
188Nemotron 3 Nano 4B NVIDIA56±6*local~3.8GB262kLM Studio › openbmb/MiniCPM5-2B model card
189LFM2.5 1.2B Instruct Liquid AI56±6*†local~2.2GB33kLM Studio › LiquidAI/LFM2.5-1.2B-Thinking model card
190Qwen3 32B Alibaba56±1api $0.13/M +local~20GB131kdeepinfra.com ↗LMArena leaderboard datasetQwen/Qwen3-Next-80B-A3B-Thinking model cardQwen/Qwen3-Next-80B-A3B-Instruct model cardArena-Hard-Auto v2.0-Preview leaderboard (repository README)ByteDance-Seed/Seed-OSS-36B-Instruct model card
191Qwen Plus (2025-01-25) Alibaba56±1*api $0.60/M131kwww.alibabacloud.com ↗LMArena leaderboard dataset
192QwQ-32B Alibaba56±1local~21GB131kLM Studio › LMArena leaderboard datasetArena-Hard-Auto v2.0-Preview leaderboard (repository README)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
193Mellum2.1 12B-A2.5B Thinking JetBrains56±8*†local~8.3GBLM Studio › JetBrains/Mellum2.1-12B-A2.5B-Thinking model card
194MiniMax-M2 MiniMax55±1*api $0.53/M +local~75GB205kdocs.aws.amazon.com ↗LMArena leaderboard dataset
195Llama 3.3 Nemotron Super 49B v1.5 NVIDIA55±1*local~18GB131kllama.cpp › LMArena leaderboard dataset
196o1-mini OpenAI55±1*comparison onlyLMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
197Mercury 2 Inception55±1*comparison onlyLMArena leaderboard dataset
198Gemma 3 12B Google55±1*api $0.075/M +local~9.1GB131kdeepinfra.com ↗LMArena leaderboard datasetGoogle DeepMind Gemma 3 270M IT model card
199Granite 4.2 30B IBM55±1*api $0.28/M +local~19GB131kdeepinfra.com ↗LMArena leaderboard datasetibm-granite/granite-4.2-30b model card
200GPT-4o (2024-05-13) OpenAI55±1api $7.50/M128kdevelopers.openai.com ↗LMArena leaderboard datasetAlpacaEval 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)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Microsoft — Phi-4-reasoning Technical Report
201Qwen3-VL 30B-A3B Instruct Alibaba55±3*†api $0.35/M +local~21GB256kwww.alibabacloud.com ↗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)
202MiniCPM5 2B OpenBMB55±6*†local~3GB131kLM Studio › openbmb/MiniCPM5-2B model card
203GPT-5 nano OpenAI55±1*api $0.14/M400kplatform.openai.com ↗LMArena leaderboard dataset
204Claude 3.5 Sonnet (2024-06-20) Anthropic55±1comparison onlyLMArena leaderboard datasetAlpacaEval 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)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
205Phi-4-reasoning 14B Microsoft54±3†api $0.22/M +local~10GB33kazure.microsoft.com ↗Microsoft — Phi-4-reasoning Technical Reportmicrosoft/Phi-4-reasoning model card
206Nova 2 Lite Amazon54±1*api $0.94/M1Mdocs.aws.amazon.com ↗LMArena leaderboard dataset
207GPT-4o (2024-08-06) OpenAI54±1api $4.38/M128kdevelopers.openai.com ↗LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
208Qwen3 8B Alibaba54±3*†api $0.31/M +local~6.1GB33kwww.alibabacloud.com ↗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)
209Qwen3 4B Instruct 2507 Alibaba54±2†local~3.9GB262kLM Studio › Qwen/Qwen3-4B-Instruct-2507 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)
210Qwen3-VL 32B Instruct Alibaba54±3*†api $0.28/M +local~22GB256kwww.alibabacloud.com ↗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)
211Llama 3.3 Nemotron Super 49B v1 NVIDIA54±1local~22GBllama.cpp › LMArena leaderboard datasetNVIDIA — Llama-Nemotron: Efficient Reasoning Models
212Nemotron Nano 9B v2 NVIDIA54±3api $0.10/M +local~7.6GB131kdocs.aws.amazon.com ↗Olmo 3NVIDIA — Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models
213Yi-Lightning 01.AI54±1*api price unknown16kplatform.lingyiwanwu.com ↗LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
214Olmo 3.1 32B Instruct Ai254±1*local~21GB66kLM Studio › LMArena leaderboard dataset
215Qwen3 4B Thinking 2507 Alibaba53±3*†local~3.9GB262kLM Studio › Qwen/Qwen3-4B-Thinking-2507 model card
216Llama 3.1 405B Instruct Meta53±1api $2.40/M +local~160GB128kdocs.aws.amazon.com ↗LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
217GPT-4 Turbo (2024-04-09) OpenAI53±1api $15.00/M128kdevelopers.openai.com ↗LMArena leaderboard datasetAlpacaEval 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)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
218Claude 3.5 Haiku Anthropic53±1*comparison onlyLMArena leaderboard datasetInception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
219Nemotron-Cascade-14B-Thinking NVIDIA53±6*†local~10GBLM Studio › NVIDIA — Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models
220Qwen3-VL 8B Instruct Alibaba53±3*†api $0.31/M +local~6.8GB256kwww.alibabacloud.com ↗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)
221Gemini 1.5 Pro (001) Google53±1*comparison onlyLMArena leaderboard dataset
222Qwen3 30B-A3B (Apr 2025, hybrid) Alibaba53±1api $0.35/M +local~19GB33kwww.alibabacloud.com ↗LMArena leaderboard datasetQwen/Qwen3-4B-Thinking-2507 model cardQwen/Qwen3-30B-A3B-Thinking-2507 model cardQwen/Qwen3-30B-A3B-Instruct-2507 model cardQwen/Qwen3-4B-Instruct-2507 model cardArena-Hard-Auto v2.0-Preview leaderboard (repository README)
223Llama 4 Maverick (402B-A17B) Meta53±1api $0.42/M +local~65GB1Mdocs.aws.amazon.com ↗LMArena leaderboard datasetArena-Hard-Auto v2.0-Preview leaderboard (repository README)
224Seed-OSS 36B Instruct ByteDance Seed53±5*†local~23GB524kLM Studio › ByteDance-Seed/Seed-OSS-36B-Instruct model card
225DeepSeek-V2.5 (1210) DeepSeek53±1*local~82GBllama.cpp › LMArena leaderboard datasetOpenCompass CompassAcademic leaderboard (full version) Space: committed predictions file
226Llama 4 Scout (109B-A17B) Meta53±1*api $0.15/M +local~36GB328kdeepinfra.com ↗LMArena leaderboard dataset
227EXAONE 4.0 32B LG AI Research53±4*†local~20GB131kLM Studio › LG AI Research vendor-reported benchmark scores for its own EXAONE models (Hugging Face model cards)
228Molmo 2 8B Ai253±1*local~6.8GBllama.cpp › LMArena leaderboard dataset
229Qwen3-VL 8B Thinking Alibaba53±3*†api $0.66/M +local~6.8GB256kwww.alibabacloud.com ↗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)
230Olmo 3 7B Think Ai253±3local~5.8GB66kLM Studio › Olmo 3JetBrains/Mellum2-12B-A2.5B-Thinking model card
231gpt-oss-20b OpenAI52±1*api $0.13/M +local~14GB131kconsole.groq.com ↗LMArena leaderboard datasetByteDance-Seed/Seed-OSS-36B-Instruct model card
232GPT-4.1 nano OpenAI52±1api $0.17/M1Mdevelopers.openai.com ↗LMArena leaderboard datasetArena-Hard-Auto v2.0-Preview leaderboard (repository README)
233Athene-V2-Chat 72B Nexusflow52±1local~44GB33kLM Studio › LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Arena-Hard-Auto v2.0-Preview leaderboard (repository README)
234Qwen3-VL 4B Thinking Alibaba52±3*†local~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)
235Gemma 3n E4B Google52±1*local~6.2GB32kllama.cpp › LMArena leaderboard dataset
236GPT-4o mini (2024-07-18) OpenAI52±1api $0.26/M128kdevelopers.openai.com ↗LMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsMicrosoft Research — Phi-4 Technical Report
237GPT-4 Turbo Preview (0125) OpenAI52±1comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
238Qwen3-VL 4B Instruct Alibaba52±3*†local~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)
239Mistral Large 2 (2407) Mistral52±1local~45GB128kllama.cpp › LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Jamba-1.5: Hybrid Transformer-Mamba Models at Scale (AI21, arXiv 2408.12570v1)
240Granite 4.0 H Micro 3B IBM52±5*†local~3.4GB131kLM Studio › ibm-granite/granite-4.0-h-small model card
241Yi-Large Preview 01.AI52±2*comparison onlyAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
242Llama 3.3 70B Instruct Meta52±1api $0.72/M +local~24GB128kdocs.aws.amazon.com ↗LMArena leaderboard datasetInception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsMicrosoft Research — Phi-4 Technical Report
243Nemotron 3 Nano 30B-A3B NVIDIA52±1*api $0.087/M +local~20GB262kdeepinfra.com ↗LMArena leaderboard dataset
244Qwen3-VL 30B-A3B Thinking Alibaba51±3*†api $0.75/M +local~21GB256kwww.alibabacloud.com ↗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)
245Athene-70B Nexusflow51±1*local~29GBllama.cpp › LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
246Olmo 3 32B Think Ai251±1local~21GB66kLM Studio › LMArena leaderboard datasetOlmo 3
247GPT-4 Turbo Preview (1106) OpenAI51±1comparison onlyLMArena leaderboard datasetAlpacaEval 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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
248Claude 3 Opus Anthropic51±1comparison onlyLMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)
249Gemini 1.5 Flash (002) Google51±1*comparison onlyLMArena leaderboard dataset
250Qwen2.5 72B Instruct Alibaba51±1api $0.37/M +local~30GB33kdeepinfra.com ↗LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsArena-Hard-Auto v2.0-Preview leaderboard (repository README)Microsoft Research — Phi-4 Technical Report
251Llama 3.1 Nemotron 70B Instruct NVIDIA51±1local~43GB128kLM Studio › LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Tülu 3: Pushing Frontiers in Open Language Model Post-Training (Lambert et al., Ai2)Arena-Hard-Auto v2.0-Preview leaderboard (repository README)
252DeepSeek-V2.5 DeepSeek51±1*local~151GB164kOllama › LMArena leaderboard datasetOpenCompass CompassAcademic leaderboard (full version) Space: committed predictions file
253Granite 4.1 8B IBM51±1*api $0.11/M +local~7GB131kreplicate.com ↗LMArena leaderboard datasetibm-granite/granite-4.1-8b model card
254Gemma 2 9B IT SimPO Princeton NLP51±1*local~7.6GBllama.cpp › LMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
255Mistral Large 2 (2411) Mistral51±1*local~78GB128kllama.cpp › LMArena leaderboard dataset
256Olmo 3 7B Instruct Ai250±3local~5.8GB66kLM Studio › Olmo 3JetBrains/Mellum2-12B-A2.5B-Instruct model card
257Magistral Medium 1.0 (2506) Mistral50±1*comparison onlyLMArena leaderboard dataset
258Gemma 3 4B Google50±1api $0.063/M +local~3.8GB131kdeepinfra.com ↗LMArena leaderboard datasetGoogle DeepMind Gemma 3 270M IT model cardLFM2 Technical Report — Liquid AI own SmolLM3 evaluation
259Granite 4.1 3B IBM50±5*†local~3.7GBOllama › ibm-granite/granite-4.1-8b model card
260Mistral Small 3.1 24B (2503) Mistral50±1*local~17GB128kllama.cpp › LMArena leaderboard datasetmistralai/Mistral-Small-3.2-24B-Instruct-2506 model card
261Granite 4.0 Micro 3B IBM50±5*†local~3.5GB131kLM Studio › ibm-granite/granite-4.0-h-small model card
262Command R7B (12-2024) Cohere49±3*local~6.2GB128kLM Studio › Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
263Flan-T5 XXL Google49±15*local~49GBhow to run › LLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusion (arXiv:2306.02561v3)
264Yi-Large 01.AI49±2*comparison onlyArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
265Qwen2.5 32B Instruct Alibaba49±2local~21GB131kLM Studio › Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
266Llama-3-Instruct-8B-SimPO-ExPO Chujie Zheng (ExPO)49±2*local~6.6GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
267Granite 4.2 3B IBM49±1api $0.052/M +local~3.7GB131kdeepinfra.com ↗LMArena leaderboard datasetibm-granite/granite-4.2-30b model cardopenbmb/MiniCPM5-2B model card
268Granite 4.2 8B IBM49±1*api $0.11/M +local~6.6GB131kwandb.ai ↗LMArena leaderboard datasetibm-granite/granite-4.2-30b model card
269Olmo 3.1 32B Think Ai249±1local~21GB66kLM Studio › LMArena leaderboard datasetOlmo 3
270Llama 3.1 Tulu 3 70B DPO Ai249±4*†local~22GB131kllama.cpp › Tülu 3: Pushing Frontiers in Open Language Model Post-Training (Lambert et al., Ai2)
271Granite 4.0 H Small (32B-A9B) IBM49±1api $0.11/M +local~21GB131kreplicate.com ↗LMArena leaderboard datasetibm-granite/granite-4.0-h-small model card
272Nova Pro Amazon49±1*api $1.40/M300kdocs.aws.amazon.com ↗LMArena leaderboard dataset
273Jamba 1.5 Large AI21 Labs49±1local~839GBhow to run › LMArena leaderboard datasetJamba-1.5: Hybrid Transformer-Mamba Models at Scale (AI21, arXiv 2408.12570v1)
274Llama-3-Instruct-8B-SimPO Princeton NLP49±2*local~6.6GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
275Granite 4.0 H Tiny (7B-A1B) IBM49±4*†local~5.6GB131kLM Studio › ibm-granite/granite-4.0-h-small model card
276Gemma 2 27B IT Google48±1local~18GB8kLM Studio › LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
277Granite 4.0 H 1B IBM48±4local~2.3GB131kLM Studio › ibm-granite/granite-4.0-h-1b model cardLiquidAI/LFM2.5-1.2B-Thinking model card
278Llama 3.1 70B Instruct Meta48±1api $0.72/M +local~27GB128kdocs.aws.amazon.com ↗LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
279Gemini 1.5 Flash (001) Google48±1*comparison onlyLMArena leaderboard dataset
280Llama 3.1 Tulu 3 70B Ai248±1local~43GB131kLM Studio › LMArena leaderboard datasetTülu 3: Pushing Frontiers in Open Language Model Post-Training (Lambert et al., Ai2)
281Llama 3.1 Nemotron 51B Instruct NVIDIA48±1*local~34GBllama.cpp › LMArena leaderboard dataset
282s1.1 32B Simple Scaling48±4*local~22GB33kLM Studio › Arena-Hard-Auto v2.0-Preview leaderboard (repository README)
283Mellum2 12B-A2.5B Instruct JetBrains48±4*†local~8.6GB131kLM Studio › JetBrains/Mellum2-12B-A2.5B-Instruct model card
284LFM2 8B-A1B Liquid AI48±4*†comparison onlyLiquidAI/LFM2.5-8B-A1B model cardLFM2 Technical Report — Liquid AI own SmolLM3 evaluation
285LFM2 2.6B Liquid AI48±6*†local~3GB33kLM Studio › LFM2 Technical Report — Liquid AI own SmolLM3 evaluation
286Qwen3 4B Alibaba47±2local~3.7GB33kLM Studio › Qwen/Qwen3-4B-Thinking-2507 model cardQwen/Qwen3-4B-Instruct-2507 model cardArena-Hard-Auto v2.0-Preview leaderboard (repository README)HuggingFaceTB/SmolLM3-3B model card (Evaluation section)LFM2 Technical Report — Liquid AI own SmolLM3 evaluation
287Ministral 3B (2410) Mistral47±6*†comparison onlyMistral AI vendor-reported benchmark scores for its own models (Hugging Face model cards)
288Granite 4.0 1B IBM47±3local~2.4GB128kLM Studio › ibm-granite/granite-4.0-h-1b model cardLiquidAI/LFM2.5-1.2B-Thinking model card
289OLMo 2 13B Instruct (1124) Allen Institute for AI47±3*†local~9.6GB4kLM Studio › 2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
290Qwen3.5 2B Alibaba47±3local~2.9GB262kllama.cpp › openbmb/MiniCPM5-2B model cardQwen/Qwen3.5-2B model cardCohereLabs/North-Micro-Vision-Instruct model card (benchmark table)
291Qwen2.5 14B Instruct Alibaba47±3local~10GB131kLM Studio › 2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsMicrosoft Research — Phi-4 Technical Report
292Nemotron-4 340B Instruct NVIDIA47±1*comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
293GLM-4 (0116) Z.ai47±2*comparison onlyArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
294GPT-4 (0314) OpenAI47±1comparison onlyLMArena leaderboard datasetAlpacaEval 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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
295Command R Plus (08-2024) Cohere47±1*api $4.38/M +local~63GB128kdocs.cohere.com ↗LMArena leaderboard dataset
296Mellum2 12B-A2.5B Thinking JetBrains47±3*†local~8.6GB131kLM Studio › JetBrains/Mellum2-12B-A2.5B-Thinking model cardJetBrains/Mellum2.1-12B-A2.5B-Thinking model card
297GLM-4 (0520) Z.ai47±1comparison onlyLMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
298Mistral Small 3 24B (2501) Mistral47±1*local~16GB32kLM Studio › LMArena leaderboard dataset
299Qwen3-VL 2B Thinking Alibaba47±3*†local~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)
300Claude 3 Sonnet Anthropic46±1comparison onlyLMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)
301Reka Core (2024-05-01) Reka AI46±2*comparison onlyPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
302Qwen2.5 Coder 32B Instruct Alibaba46±1*api $0.74/M +local~21GB131kdevelopers.cloudflare.com ↗LMArena leaderboard dataset
303Aya Expanse 32B Cohere46±1*api $0.75/M +local~22GB128kdocs.cohere.com ↗LMArena leaderboard datasetInception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
304North Micro Vision Instruct (2.4B) Cohere46±8*†local~3.8GB128khow to run › CohereLabs/North-Micro-Vision-Instruct model card (benchmark table)
305Mistral Small 4 (119B-A6.5B) Mistral46±7*†api $0.26/M +local~40GB262kdocs.mistral.ai ↗mistralai/Mistral-Small-4-119B-2603 model card
306Gemma 2 9B IT Google46±1local~7GB8kLM Studio › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsJamba-1.5: Hybrid Transformer-Mamba Models at Scale (AI21, arXiv 2408.12570v1)
307Llama 3 70B Instruct Meta46±1local~24GB8kllama.cpp › LMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)
308DeepSeek-Coder-V2 Instruct (0614) DeepSeek46±1local~92GBllama.cpp › LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
309Qwen2.5 7B Instruct Alibaba45±2api $0.30/M +local~6GB131kwww.together.ai ↗Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
310Claude 3 Haiku Anthropic45±1comparison onlyLMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
311OpenThinker2 32B Open Thoughts45±5*local~21GB33kLM Studio › Arena-Hard-Auto v2.0-Preview leaderboard (repository README)
312GPT-4 (0613) OpenAI45±1api $37.50/M8kdevelopers.openai.com ↗LMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
313Qwen2 72B Instruct Alibaba45±1local~33GB131kllama.cpp › LMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)
314Nova Lite Amazon45±1*api $0.10/M300kdocs.aws.amazon.com ↗LMArena leaderboard dataset
315Granite 3.3 8B Instruct IBM45±4*†local~6.3GB128kLM Studio › ibm-granite/granite-3.3-8b-instruct model card
316LFM2 1.2B Liquid AI45±4*†local~2.2GB33kLM Studio › LiquidAI/LFM2-1.2B model card
317Gemini 1.5 Flash-8B (001) Google45±1*comparison onlyLMArena leaderboard dataset
318Command R Plus (04-2024) Cohere44±1local~43GB128kllama.cpp › LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
319Phi-4 14B Microsoft44±1api $0.087/M +local~10GB16kdeepinfra.com ↗LMArena leaderboard datasetMicrosoft — Phi-4-reasoning Technical ReportMicrosoft Research — Phi-4 Technical Reportmicrosoft/Phi-4-reasoning model card
320OLMo 2 32B Instruct (0325) Ai244±1local~21GBLM Studio › LMArena leaderboard dataset2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
321Granite 3.2 8B Instruct IBM44±4*†local~6.6GB131kllama.cpp › ibm-granite/granite-3.3-8b-instruct model card
322EXAONE 3.5 2.4B Instruct LG AI Research44±3*†local~3GB33kLM Studio › LG AI Research vendor-reported benchmark scores for its own EXAONE models (Hugging Face model cards)
323Qwen3 1.7B Alibaba44±2local~2.6GB33kLM Studio › LiquidAI/LFM2-1.2B model cardQwen/Qwen3.5-2B model cardHuggingFaceTB/SmolLM3-3B model card (Evaluation section)LiquidAI/LFM2.5-1.2B-Thinking model card
324Qwen2.5 3B Instruct Alibaba44±3*local~3.3GB33kLM Studio › Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
325Qwen2 7B Instruct Alibaba44±3*local~6GB131kLM Studio › Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
326Command R (08-2024) Cohere44±1*api $0.26/M +local~21GB128kdocs.cohere.com ↗LMArena leaderboard dataset
327Ministral 3 14B Mistral43±3*api $0.20/M +local~9.1GB262kdocs.mistral.ai ↗JetBrains/Mellum2-12B-A2.5B-Instruct model cardmistralai/Ministral-3-14B-Instruct-2512 model card
328LFM2 700M Liquid AI43±4*†local~1.9GB33kLM Studio › LiquidAI/LFM2-1.2B model card
329Qwen1.5 110B Chat Alibaba43±1local~45GB32kllama.cpp › LMArena leaderboard datasetAlpacaEval 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)
330Qwen3-VL 2B Instruct Alibaba43±2*local~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)
331Nova Micro Amazon42±1*api $0.061/M128kdocs.aws.amazon.com ↗LMArena leaderboard dataset
332Mistral Large (2402) Mistral42±1comparison onlyLMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)
333Jamba 1.5 Mini AI21 Labs42±1local~34GBllama.cpp › LMArena leaderboard datasetJamba-1.5: Hybrid Transformer-Mamba Models at Scale (AI21, arXiv 2408.12570v1)
334SmolLM3 3B Hugging Face42±4*local~3.3GB66kLM Studio › HuggingFaceTB/SmolLM3-3B model card (Evaluation section)LFM2 Technical Report — Liquid AI own SmolLM3 evaluation
335EXAONE 4.0 1.2B LG AI Research42±3*†local~2.2GB66kLM Studio › LG AI Research vendor-reported benchmark scores for its own EXAONE models (Hugging Face model cards)
336Gemini 1.0 Pro (001) Google42±1*comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
337Ministral 8B (2410) Mistral42±1local~6.2GB128kLM Studio › LMArena leaderboard dataset2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsMistral AI vendor-reported benchmark scores for its own models (Hugging Face model cards)
338Ministral 3 8B Mistral42±6*†api $0.15/M +local~6.5GB262kdocs.mistral.ai ↗mistralai/Ministral-3-14B-Instruct-2512 model card
339OLMo 2 7B Instruct (1124) Allen Institute for AI42±4*†local~5.8GB4kLM Studio › 2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
340Qwen1.5 72B Chat Alibaba41±1local~31GB32kllama.cpp › LMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
341Mixtral 8x22B Instruct v0.1 Mistral41±1local~56GB64kllama.cpp › LMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)
342Aya Expanse 8B Cohere41±1*local~6.2GB8kLM Studio › LMArena leaderboard datasetInception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
343Reka Flash 21B (2024-02-26) Reka AI41±1*comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
344Mistral Medium (2312) Mistral40±1comparison onlyLMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)
345Llama 3.1 Tulu 3 8B Ai240±1local~6.2GB131kLM Studio › LMArena leaderboard dataset2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
346GPT-3.5 Turbo (0125) OpenAI40±1api $0.75/M16kdevelopers.openai.com ↗LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
347Command R (03-2024) Cohere40±1local~24GB128kllama.cpp › LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
348Claude Instant 1.2 Anthropic40±2*comparison onlyAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
349Claude 2.0 Anthropic39±1comparison onlyAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
350Xwin-LM 70B v0.1 Xwin-LM39±2*local~28GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
351Llama 3 8B Instruct Meta39±1local~6.6GB8kllama.cpp › LMArena leaderboard datasetAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)
352GPT-3.5 Turbo (0613) OpenAI39±1comparison onlyAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
353LFM2 350M Liquid AI39±4*†local~1.7GB33kLM Studio › LiquidAI/LFM2-1.2B model card
354Zephyr ORPO 141B-A35B v0.1 Hugging Face39±1local~56GB66kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)HuggingFaceH4 zephyr-orpo-141b-A35b-v0.1 model card
355Llama 3.1 8B Instruct Meta39±1api $0.22/M +local~6.2GB128kdocs.aws.amazon.com ↗LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsIBM Granite 3.0 Language Models technical report (GitHub)
356Granite 3.1 8B Instruct IBM39±1local~6.6GB128kllama.cpp › LMArena leaderboard datasetibm-granite/granite-3.3-8b-instruct model cardibm-granite/granite-3.1-3b-a800m-instruct model card
357Yi 1.5 34B Chat 01.AI39±1*local~22GB4kLM Studio › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
358Claude 2.1 Anthropic38±1comparison onlyAlpacaEval 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)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
359Gemma 3 1B Google38±2local~2.2GB32kLM Studio › 2 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 cardGoogle DeepMind Gemma 3 270M IT model card
360Qwen3 0.6B Alibaba38±3*local~1.8GB33kLM Studio › LiquidAI/LFM2-1.2B model cardOlmo 3
361Ministral 3 14B Reasoning 2512 Mistral38±3*local~10GB262kllama.cpp › JetBrains/Mellum2-12B-A2.5B-Thinking model card
362Qwen1.5 32B Chat Alibaba38±1*local~22GB32kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
363Phi-3-medium-4k-instruct Microsoft38±1local~9.8GB4kLM Studio › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
364Gemma 2 2B IT Google37±1*local~3.1GB8kLM Studio › LMArena leaderboard datasetIBM Granite 3.0 Language Models technical report (GitHub)
365Phi-3.5-vision-instruct Microsoft37±8*local~10GBhow to run › CohereLabs/North-Micro-Vision-Instruct model card (benchmark table)
366Mixtral 8x7B Instruct v0.1 Mistral37±1local~29GB32kLM Studio › LMArena leaderboard datasetAlpacaEval 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)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)HuggingFaceH4 zephyr-orpo-141b-A35b-v0.1 model card
367GPT-3.5 Turbo (1106) OpenAI37±1comparison onlyLMArena leaderboard datasetAlpacaEval 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)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
368DBRX Instruct Databricks37±1local~85GB32kOllama › LMArena leaderboard datasetAlpacaEval 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)HuggingFaceH4 zephyr-orpo-141b-A35b-v0.1 model card
369InternLM2.5 20B Chat Shanghai AI Laboratory (InternLM)37±1*local~14GB33kllama.cpp › LMArena leaderboard datasetArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
370InternLM2 Chat 20B Shanghai AI Laboratory (InternLM)36±2*local~14GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
371Mistral 7B Instruct v0.3 Mistral36±2*local~5.8GB32kLM Studio › AlpacaEval 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)
372Yi 34B Chat 01.AI36±1local~23GB4kllama.cpp › LMArena leaderboard datasetAlpacaEval 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)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
373Qwen1.5 14B Chat Alibaba36±1local~11GB32kOllama › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
374Granite 3.3 2B Instruct IBM36±3*†local~3GB128kLM Studio › ibm-granite/granite-3.3-8b-instruct model card
375WizardLM 70B V1.0 Microsoft35±1local~32GBllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
376OpenChat 3.5 0106 OpenChat35±1*local~6.1GB8kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
377Tulu 2 DPO 70B Ai235±1local~32GBllama.cpp › LMArena leaderboard datasetAlpacaEval 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)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
378Granite 3.1 3B-A800M Instruct IBM35±5*†local~3.6GB131kllama.cpp › ibm-granite/granite-3.1-3b-a800m-instruct model card
379Snowflake Arctic Instruct Snowflake35±1local~1007GBhow to run › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
380Granite 3.0 8B Instruct IBM35±1*local~6.6GB4kllama.cpp › LMArena leaderboard datasetIBM Granite 3.0 Language Models technical report (GitHub)
381OpenChat 3.5 OpenChat35±1*local~6.1GB8kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
382DeepSeek LLM 67B Chat DeepSeek35±1local~31GB4kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
383GPT-3.5 Turbo (0301) OpenAI35±2comparison onlyAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
384Gemma 4 E4B Google35±2local~6.9GB131kLM Studio › XHToken/Spark-X2.5-4B model cardopenbmb/MiniCPM5-2B model cardJetBrains/Mellum2.1-12B-A2.5B-Thinking model cardGemma Team — Gemma 4 Technical Report
385Gemma 1.1 7B IT Google35±1local~7.1GB8kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
386Granite 3.1 2B Instruct IBM35±1local~3.1GB128kllama.cpp › LMArena leaderboard datasetibm-granite/granite-3.3-8b-instruct model cardibm-granite/granite-3.1-3b-a800m-instruct model card
387Phi-3-small-8k-instruct Microsoft34±1local~17GBhow to run › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
388Starling-LM 7B beta Nexusflow34±1local~6GB8kLM Studio › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
389Ministral 3 3B Mistral34±5*api $0.10/M +local~3.5GB262kdocs.mistral.ai ↗mistralai/Ministral-3-14B-Instruct-2512 model cardCohereLabs/North-Micro-Vision-Instruct model card (benchmark table)
390Deita 7B v1.0 HKUST NLP34±2*local~32GBhow to run › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
391Granite 4.0 H 350M IBM34±4*†local~1.7GB33kLM Studio › ibm-granite/granite-4.0-h-1b model card
392Vicuna 33B v1.3 LMSYS34±1local~23GB2kOllama › LMArena leaderboard datasetAlpacaEval 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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard tableArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
393OpenHermes 2.5 Mistral 7B Nous Research34±1local~6.1GBllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
394Granite 3.2 2B Instruct IBM34±3*†local~3GB131kLM Studio › ibm-granite/granite-3.3-8b-instruct model card
395Llama 3.2 3B Instruct Meta33±1api $0.12/M +local~3.4GB131kdevelopers.cloudflare.com ↗LMArena leaderboard datasetIBM Granite 3.0 Language Models technical report (GitHub)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)LFM2 Technical Report — Liquid AI own SmolLM3 evaluation
396Llama 2 70B Chat Meta33±1local~32GB4kllama.cpp › LMArena leaderboard datasetAlpacaEval 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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard tableArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
397OLMo 2 1B Instruct (0425) Allen Institute for AI33±4*†local~2.4GB4kllama.cpp › 2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct models
398Starling-LM 7B alpha UC Berkeley (NEST)33±1local~6.1GB8kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
399OpenThinker3 7B Open Thoughts33±3*local~6GB33kLM Studio › Olmo 3
400Nous Hermes 2 Mixtral 8x7B DPO Nous Research33±1*local~31GB33kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
401Gemma 3 270M IT Google32±5*†local~1.7GB32kLM Studio › Google DeepMind Gemma 3 270M IT model card
402Granite 3.0 2B Instruct IBM32±1*local~3.2GB4kllama.cpp › LMArena leaderboard datasetIBM Granite 3.0 Language Models technical report (GitHub)
403SmolLM2 360M Instruct Hugging Face32±5*†local~1.7GB8kLM Studio › HuggingFaceTB/SmolLM2-360M-Instruct model card
404QwQ-32B-Preview Alibaba31±1*local~21GB33kLM Studio › LMArena leaderboard dataset
405WizardLM 13B V1.2 Microsoft31±1local~9.8GBllama.cpp › LMArena leaderboard datasetAlpacaEval 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)
406Mistral 7B Instruct v0.2 Mistral31±1local~6GB32kLM Studio › LMArena leaderboard datasetAlpacaEval 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)
407Tulu 2 DPO 13B Allen Institute for AI31±2*local~9.8GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
408Granite 3.1 1B-A400M Instruct IBM31±5*†local~2.3GB131kllama.cpp › ibm-granite/granite-3.1-3b-a800m-instruct model card
409Granite 3.0 3B-A800M Instruct IBM31±4*†local~3.7GB4kllama.cpp › IBM Granite 3.0 Language Models technical report (GitHub)
410SOLAR 10.7B Instruct v1.0 Upstage31±1*local~8.3GB4kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
411Qwen3.5 0.8B Alibaba31±4*†local~2GB262kllama.cpp › Qwen/Qwen3.5-2B model card
412Dolphin 2.2.1 Mistral 7B Cognitive Computations31±1local~6.1GBllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
413Qwen1.5 7B Chat Alibaba31±1local~6.5GB32kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
414Falcon-180B-Chat TII30±1*local~115GB2kOllama › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
415Qwen2.5 0.5B Instruct Alibaba30±3*local~1.8GB33kLM Studio › Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)HuggingFaceTB/SmolLM2-360M-Instruct model card
416MPT-30B-Chat MosaicML30±1comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
417Qwen-14B-Chat Alibaba30±1local~11GBllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
418Phi-3-mini-4k-instruct (June 2024 update) Microsoft30±1*local~3.7GBLM Studio › LMArena leaderboard dataset
419Vicuna 13B v1.5 LMSYS30±1local~9.8GB4kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard tableAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
420Granite 4.0 350M IBM30±4*†local~1.7GB32kLM Studio › ibm-granite/granite-4.0-h-1b model card
421Llama 2 13B Chat Meta29±1local~9.8GB4kllama.cpp › LMArena leaderboard datasetAlpacaEval 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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard tableInception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
422Code Llama 34B Instruct Meta29±1*local~22GB16kLM Studio › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
423Zephyr 7B beta Hugging Face29±1local~6.1GBllama.cpp › LMArena leaderboard datasetAlpacaEval 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)
424Tulu 2 DPO 7B Allen Institute for AI29±2*local~5.8GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
425Phi-3-mini-128k-instruct (Apr 2024) Microsoft29±1local~9.5GBhow to run › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
426Gemma 7B IT Google29±1local~7.1GB8kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
427Phi-3-mini-4k-instruct (Apr 2024) Microsoft29±1*local~4GBllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
428Zephyr 7B alpha Hugging Face28±1local~6.1GBllama.cpp › LMArena leaderboard datasetAlpacaEval 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)
429PaLM 2 Chat (chat-bison@001) Google28±1comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
430Gemma 4 E2B Google28±3api $0.050/M +local~5.1GB131kdocs.aws.amazon.com ↗XHToken/Spark-X2.5-4B model cardopenbmb/MiniCPM5-2B model cardCohereLabs/North-Micro-Vision-Instruct model card (benchmark table)Gemma Team — Gemma 4 Technical Report
431Llama 3.2 1B Instruct Meta27±1api $0.020/M +local~2.2GB131knovita.ai ↗LMArena leaderboard datasetLiquidAI/LFM2-1.2B model cardIBM 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/LFM2.5-1.2B-Thinking model cardSmolLM2 paper (Hugging Face) — authors' own instruct-model evaluations+1
432Code Llama 70B Instruct Meta27±1*local~28GB16kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
433Guanaco 33B University of Washington27±1local~70GBhow to run › LMArena leaderboard datasetAlpacaEval 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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
434Vicuna 7B v1.5 LMSYS27±1local~5.8GB4kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard tableAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
435Gemma 1.1 2B IT Google27±1local~3.2GB8kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
436SmolLM2 1.7B Instruct Hugging Face27±1local~2.6GB8kLM Studio › LMArena leaderboard dataset2 OLMo 2 Furious (Team OLMo, Ai2) — authors' own evaluations of OLMo 2 Instruct and peer instruct modelsSmolLM2 paper (Hugging Face) — authors' own instruct-model evaluations
437Granite 3.0 1B-A400M Instruct IBM26±4*†local~2.3GB4kllama.cpp › IBM Granite 3.0 Language Models technical report (GitHub)
438Qwen2.5 1.5B Instruct Alibaba26±3local~2.4GB33kLM Studio › 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 evaluations
439Mistral 7B Instruct v0.1 Mistral26±1*local~6.1GB8kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
440Llama 2 7B Chat Meta25±1local~5.8GB4kllama.cpp › LMArena leaderboard datasetAlpacaEval 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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard tableArena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)
441Ghost 7B Alpha Ghost X25±2*local~6.1GBllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
442Nous-Hermes-13b Nous Research25±2*local~9.8GB2kllama.cpp › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
443Qwen1.5 4B Chat Alibaba24±1*local~4.1GB32kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
444Gemma 2B IT Google23±1local~3.1GB8kllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)
445OLMo 7B Instruct Ai221±1*local~16GBhow to run › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
446Koala 13B UC Berkeley (BAIR)21±1comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
447Alpaca 13B Stanford CRFM20±1comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
448GPT4All-13B-Snoozy Nomic AI20±1local~56GBhow to run › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
449MPT-7B-Chat MosaicML20±1comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
450Falcon 40B Instruct TII19±1local~27GBOllama › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
451ChatGLM3-6B Zhipu AI (THUDM)19±1*local~5.8GBllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)
452Falcon 7B Instruct TII18±2*local~5.9GBOllama › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
453RWKV-4-Raven 14B RWKV (BlinkDL)17±1comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
454ChatGLM2-6B Zhipu AI (THUDM)16±1local~15GBhow to run › LMArena leaderboard datasetAlpacaEval 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)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
455Qwen1.5 1.8B Chat Alibaba16±5*local~2.7GB33kOllama › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
456Phi-2 (2.7B) Microsoft14±2*local~3.2GB2kLM Studio › AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
457OpenAssistant SFT-4 Pythia 12B LAION (OpenAssistant)14±1local~9.5GBllama.cpp › LMArena leaderboard datasetAlpacaEval 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)LMSYS 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)
458GPT-3 text-davinci-001 OpenAI12±2*comparison onlyAlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)
459ChatGLM-6B Zhipu AI (THUDM)12±1local~16GBhow to run › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS 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)
460FastChat-T5 3B LMSYS10±1local~8.5GBhow to run › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
461Dolly v2 12B Databricks9±1comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS 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)
462LLaMA 13B Meta8±1comparison onlyLMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table
463StableLM-Tuned-Alpha 7B Stability AI6±1local~6.9GBllama.cpp › LMArena leaderboard datasetPolyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)LMSYS 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)
—ERNIE 4.5 300B-A47B Baidunot yet calibratedlocal~97GB131kllama.cpp › ERNIE 4.5 post-trained 300B-A47B benchmark chart
—Hermes 3 Llama 3.1 405B Nous Researchnot yet calibratedlocal~114GB131kllama.cpp › Nous Research vendor-reported scores for its own Hermes models (Hermes 3 Technical Report)
—Hermes 3 Llama 3.1 70B Nous Researchnot yet calibratedlocal~44GB131kOllama › Nous Research vendor-reported scores for its own Hermes models (Hermes 3 Technical Report)
—Hermes 3 Llama 3.1 8B Nous Researchnot yet calibratedlocal~6.6GB128kllama.cpp › Nous Research vendor-reported scores for its own Hermes models (Hermes 3 Technical Report)
—Phi-3.5-mini-instruct Microsoftnot yet calibratedlocal~3.7GB128kLM Studio › Microsoft — Phi-4-Mini Technical Report
—Phi-4-mini 3.8B Microsoftnot yet calibratedapi $0.13/M +local~3.8GB128kazure.microsoft.com ↗Microsoft — Phi-4-Mini Technical Report
—Qwen2 0.5B Instruct Alibabanot yet calibratedlocal~1.8GB32kLM Studio › Qwen2-0.5B-Instruct model card
—Qwen2 1.5B Instruct Alibabanot yet calibratedlocal~2.4GB32kLM Studio › Qwen2-0.5B-Instruct model card

+ 116 models with fewer than 3 results left off, since one or two tests say little about overall ability · show them

Includes 113 comparison-only models with no verified API or download · hide them

471 of 588 catalogue models listed here · sources and licences

Models outside this leaderboard (1 models)

No current run option verified: ERNIE 4.5 21B-A3B

* 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