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

Best AI models for images and documents for a 32GB graphics card

A leaderboard of the best AI models for images and documents for a 32GB graphics card (with 32GB of system RAM). The best one that runs on it right now is Qwen3.8 27B, followed by Gemma 4 31B and Qwen3.6 27B. Models are ranked on tests that ask questions about images, charts and documents, like MMMU and CharXiv. Models that run on it show the download that fits and the command to run it.

140 models · scores combine 47 benchmarks · updated how it works · find a leaderboard · benchmark your machine
task[all tasks][agent tasks][coding][chat & writing][math][vision & documents][web search]

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

Models for 32GB RAM · 32GB graphics card → ~32GB fast · reset
+ more
card[RTX 5090][RTX 5000 Ada][RTX PRO 4500 Blackwell]
set[all sets][Arena vision (style controlled) · 199][CharXiv descriptive questions · 2]

Filter the model list while keeping task scores. Models without a shared score scale are not yet calibrated; select a set to see its own scores.

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=32&gpu=card&vram=32"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]
vram[1][2][4][8][12][16][24][32][48][64][80][96][128][160]GB
price[local][$0.05][$0.25][$1][$5][$20][any]max per M

detecting your OS…

Score vs API price

#modeltask score↓where↓speed↓RAM↓context↓get itsources
#modeltask score↓where↓speed↓RAM↓context↓get itsources
1Claude Mythos Preview (limited preview) Anthropic73±6*†limited preview1MAnthropic system cards and model announcements (lab-reported scores)
2Qwen3.8 Max (2.4T-A95B) Alibaba71±2*api $3.00/M~380GB262kdeepinfra.com ↗LMArena leaderboard datasetQwen (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)
3Claude Fable 5 Anthropic71±2api $20.00/M1Mdocs.anthropic.com ↗LMArena leaderboard datasetMeridian Labs inspect_charxiv evaluation report (GitHub README)
4Claude Opus 4.7 Anthropic70±2api $10.00/M1Mdocs.anthropic.com ↗LMArena leaderboard datasetAnthropic system cards and model announcements (lab-reported scores)
5Gemini 3.7 Flash Google69±2*api $1.50/M1Mai.google.dev ↗LMArena leaderboard datasetGoogle DeepMind Gemini model page and model cards (lab-reported scores)
6Claude Opus 5 Anthropic69±2api $10.00/M1Mdocs.anthropic.com ↗LMArena leaderboard datasetSciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
7Gemini 3.8 Flash Google69±2*api $1.50/M1Mai.google.dev ↗LMArena leaderboard datasetGoogle DeepMind Gemini model page and model cards (lab-reported scores)
8Muse Spark 1.2 Meta69±2*api $2.00/M1Mdev.meta.ai ↗LMArena leaderboard dataset
9GPT-6 Astra OpenAI69±2api $20.00/M1.1Mplatform.openai.com ↗LMArena leaderboard datasetMeridian Labs inspect_charxiv evaluation report (GitHub README)OpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)
11GPT-5.6 Sol OpenAI69±2api $8.00/M1.1Mplatform.openai.com ↗LMArena leaderboard datasetSciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)OpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
12Claude Fable 5.1 Anthropic68±2*api $20.00/M1Mdocs.anthropic.com ↗LMArena leaderboard dataset
13GPT-6.1 Sol OpenAI68±2*api $4.00/M1.1Mdevelopers.openai.com ↗LMArena leaderboard dataset
15Qwen3.8-Omni-Flash Alibaba68±4*†api $0.23/M1Mwww.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)
16Claude Opus 4.8 Anthropic68±2api $10.00/M1Mdocs.anthropic.com ↗LMArena leaderboard datasetSciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)Anthropic system cards and model announcements (lab-reported scores)
17Gemini 3.5 Flash Google68±2api $3.38/M1Mai.google.dev ↗LMArena leaderboard datasetGoogle DeepMind Gemini model page and model cards (lab-reported scores)SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)
18Muse Spark 1.3 Meta68±2*api $2.00/M1Mdev.meta.ai ↗LMArena leaderboard dataset
19Claude Opus 4.6 Anthropic68±2api $10.00/M1Mdocs.anthropic.com ↗LMArena leaderboard datasetAnthropic system cards and model announcements (lab-reported scores)BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)Diagram-MMU: A Multi-Modal Benchmark for Scientific Diagrams (arXiv:2608.12262v1)
20Gemini 3.1 Pro (Preview) Google67±2api $4.50/M1Mai.google.dev ↗LMArena leaderboard datasetGoogle DeepMind Gemini model page and model cards (lab-reported scores)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)Diagram-MMU: A Multi-Modal Benchmark for Scientific Diagrams (arXiv:2608.12262v1)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
21Gemini 3.6 Flash Google67±2api $1.50/M1Mai.google.dev ↗LMArena leaderboard datasetGoogle DeepMind Gemini model page and model cards (lab-reported scores)SciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
22GPT-5.5 OpenAI67±2api $11.25/M1.1Mplatform.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)
23Muse Spark 1.1 Meta67±2*api $2.00/M1Mdev.meta.ai ↗LMArena leaderboard dataset
25GLM-5.3-Flash Z.ai67±2*api $0.24/M~103GB1Mdocs.z.ai ↗LMArena leaderboard datasetHow Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
26Kimi K3 Moonshot AI67±6*†api $6.00/M~884GB1Mplatform.moonshot.ai ↗Moonshot AI vendor-reported benchmark scores for its own Kimi models (Hugging Face model cards)
27GPT-5.4 OpenAI67±2api $5.63/M1.1Mplatform.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)
28Grok 4.5 SpaceXAI (xAI)67±2*api $3.00/M500kdocs.x.ai ↗LMArena leaderboard dataset
29Qwen3.7 Plus Alibaba66±2api $0.70/M1Mwww.alibabacloud.com ↗LMArena leaderboard datasetQwen (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)Qwen/Qwen3.8-27B model cardSciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)
31Qwen3.6 Plus Alibaba66±4*api $1.13/M1Mdocs.together.ai ↗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)BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)
32Gemini 3 Flash (Preview) Google66±2api $1.13/M1Mai.google.dev ↗LMArena leaderboard datasetGoogle DeepMind Gemini model page and model cards (lab-reported scores)SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)Diagram-MMU: A Multi-Modal Benchmark for Scientific Diagrams (arXiv:2608.12262v1)
33GPT-5.6 Terra OpenAI66±2api $4.50/M1.1Mplatform.openai.com ↗LMArena leaderboard datasetSciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)OpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)
34Claude Sonnet 4.6 Anthropic65±2api $6.00/M1Mdocs.anthropic.com ↗LMArena leaderboard datasetSciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)Anthropic system cards and model announcements (lab-reported scores)SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)VisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback? (arXiv:2608.10408v1)BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)
35Claude Sonnet 5.5 Anthropic65±2*api $4.00/M1Mplatform.claude.com ↗LMArena leaderboard dataset
36Claude Sonnet 5 Anthropic65±2api $4.00/M1Mdocs.anthropic.com ↗LMArena leaderboard datasetAnthropic system cards and model announcements (lab-reported scores)SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)
37Qwen3.6 Flash Alibaba65±6*api $0.56/M1Mwww.alibabacloud.com ↗SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)
38Grok 4.6 SpaceXAI (xAI)65±2*api $3.00/M500kdocs.x.ai ↗LMArena leaderboard dataset
39Qwen3.5 397B-A17B Alibaba65±2api $1.09/M~65GB262kdeepinfra.com ↗LMArena leaderboard datasetQwen/Qwen3.5-397B-A17B 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)BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)Diagram-MMU: A Multi-Modal Benchmark for Scientific Diagrams (arXiv:2608.12262v1)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
40Kimi K2.6 Moonshot AI64±2api $1.71/M~317GB262kplatform.moonshot.ai ↗LMArena leaderboard datasetMoonshot AI vendor-reported benchmark scores for its own Kimi models (Hugging Face model cards)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
41GPT-5.6 Luna OpenAI64±2api $0.45/M1.1Mplatform.openai.com ↗LMArena leaderboard datasetSciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)OpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)
42Qwen3.8 27B · 4-bit Alibaba64±2local~22GB262kllama.cpp › LMArena leaderboard datasetQwen (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)Qwen/Qwen3.8-27B model cardSciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
44Gemini 3.5 Flash-Lite Google64±2*api $0.85/M1Mai.google.dev ↗LMArena leaderboard datasetGoogle DeepMind Gemini model page and model cards (lab-reported scores)
46Grok 4.20 (0309 reasoning) SpaceXAI (xAI)63±2*api $1.56/M1Mdocs.x.ai ↗LMArena leaderboard dataset
47Kimi K2.5 Moonshot AI63±2api $1.20/M~160GB262kdocs.aws.amazon.com ↗LMArena leaderboard datasetSciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)Qwen/Qwen3.5-397B-A17B model cardChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)Moonshot AI vendor-reported benchmark scores for its own Kimi models (Hugging Face model cards)Diagram-MMU: A Multi-Modal Benchmark for Scientific Diagrams (arXiv:2608.12262v1)
48GPT-5.4 mini OpenAI62±2*api $1.69/M400kplatform.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)
49GPT-5.2 OpenAI62±2api $4.81/M400kplatform.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)Qwen/Qwen3.5-397B-A17B model cardDiagram-MMU: A Multi-Modal Benchmark for Scientific Diagrams (arXiv:2608.12262v1)
50Grok 4.20 Multi-Agent (0309) SpaceXAI (xAI)62±2*api $1.56/M1Mdocs.x.ai ↗LMArena leaderboard dataset
51MiMo-V2.6-Flash (309B-A15B) Xiaomi62±2*api $0.17/M~136GB1Mmimo.mi.com ↗LMArena leaderboard dataset
52Gemma 4 31B · 4-bit Google62±2local~21GB262kOllama › LMArena leaderboard datasetGemma Team — Gemma 4 Technical Reportmeta-models/Muse-Glimmer-30B model card (speed and benchmark tables)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)
53Qwen3.6 27B · 4-bit Alibaba62±2local~22GB262kllama.cpp › Qwen/Qwen3.6-27B model cardQwen/Qwen3.8-27B model cardmeta-models/Muse-Glimmer-30B model card (speed and benchmark tables)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
54GPT-5.1 OpenAI62±2api $3.44/M400kplatform.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)
55MiMo-V2.6-Pro Xiaomi62±2*api $0.54/M~443GB1Mmimo.mi.com ↗LMArena leaderboard dataset
56Qwen3.5 35B-A3B · 4-bit Alibaba61±2local~25GB262kOllama › Qwen/Qwen3.5-122B-A10B model cardQwen/Qwen3.6-35B-A3B model cardChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)Qwen/Qwen3.5-27B model cardHow Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
57Grok 4.3 SpaceXAI (xAI)61±2*api $1.56/M1Mdocs.x.ai ↗LMArena leaderboard dataset
58Qwen3.5 122B-A10B Alibaba61±2api $1.10/M~79GB262kwww.alibabacloud.com ↗Qwen/Qwen3.5-122B-A10B model cardLMArena leaderboard datasetQwen/Qwen3.5-27B model cardHow Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
59Gemini 2.5 Pro Google61±2api $3.44/M1Mai.google.dev ↗LMArena leaderboard datasetGoogle DeepMind Gemini model page and model cards (lab-reported scores)ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)Baidu - ERNIE 5.0 Technical Report (arXiv:2602.04705v1)BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (Ai2, arXiv 2601.10611v4)
60Qwen3.6 35B-A3B · 4-bit Alibaba61±3local~24GB262kOllama › Qwen/Qwen3.6-27B model cardQwen/Qwen3.6-35B-A3B model cardHow Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
61Gemma 4 26B-A4B · 8-bit Google61±2local~30GB262kllama.cpp › LMArena leaderboard datasetGemma Team — Gemma 4 Technical Reportgoogle/diffusiongemma-26B-A4B-it model card
62Claude Opus 4.5 Anthropic61±3api $10.00/M200kdocs.anthropic.com ↗LMArena leaderboard datasetAnthropic system cards and model announcements (lab-reported scores)Qwen/Qwen3.5-397B-A17B model card
64Qwen3.5 27B · 4-bit Alibaba60±1local~20GB262kOllama › Qwen/Qwen3.5-122B-A10B model cardLMArena leaderboard datasetQwen/Qwen3.6-35B-A3B model cardQwen/Qwen3.5-27B model cardSpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)BEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)+1
65MiniMax-M3 MiniMax60±2api $0.48/M~137GB1Mdeepinfra.com ↗LMArena leaderboard datasetHow Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
66Gemini 3.1 Flash-Lite Google60±2*api $0.56/M1Mai.google.dev ↗LMArena leaderboard datasetGoogle DeepMind Gemini model page and model cards (lab-reported scores)
67Claude Haiku 4.5 Anthropic60±4*api $2.00/M200kdocs.anthropic.com ↗LMArena leaderboard dataset
68Muse Glimmer 30B · 4-bit Meta60±4†local~21GB131kOllama › meta-models/Muse-Glimmer-30B model card (speed and benchmark tables)
69MiMo-V2.5 (310B-A15B) Xiaomi60±2api $0.21/M~100GB1Mnovita.ai ↗LMArena leaderboard datasetSciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)XiaomiMiMo/MiMo-V2.5 model cardXiaomi MiMo vendor-reported benchmark scores for its own models (mimo.xiaomi.com launch pages)
71GLM-5V-Turbo Z.ai59±2*api $1.90/M200kwww.siliconflow.com ↗LMArena leaderboard dataset
72Inkling (975B-A41B) Thinking Machines59±5*†api $1.73/M~156GB1Mdeepinfra.com ↗thinkingmachines/Inkling-Small model card
74Claude Sonnet 4.5 Anthropic58±3api $6.00/M200kdocs.anthropic.com ↗LMArena leaderboard datasetVisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback? (arXiv:2608.10408v1)Qwen/Qwen3.5-27B model cardMolmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (Ai2, arXiv 2601.10611v4)
75o3 OpenAI57±2*api $3.50/M200kplatform.openai.com ↗LMArena leaderboard dataset
77GPT-5 OpenAI57±2api $3.44/M400kplatform.openai.com ↗LMArena leaderboard datasetServiceNow-AI/Apriel-1.6-15b-Thinker model cardBaidu - ERNIE 5.0 Technical Report (arXiv:2602.04705v1)VisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback? (arXiv:2608.10408v1)ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (Ai2, arXiv 2601.10611v4)
78Qwen3.5 9B · 8-bit Alibaba57±2local~12GB262kllama.cpp › Qwen/Qwen3.5-9B model cardChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)Qwen/Qwen3.5-4B model cardBEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
79Inkling Small (266B-A12B) Thinking Machines56±2api $0.64/M~86GB1Mdeepinfra.com ↗LMArena leaderboard datasetthinkingmachines/Inkling-Small model card
80Gemma 4 12B · 8-bit Google56±4†local~15GB262kllama.cpp › Gemma Team — Gemma 4 Technical Report
83GPT-4.1 OpenAI56±2api $3.50/M1Mplatform.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)
85o4-mini OpenAI55±2*api $1.93/M200kplatform.openai.com ↗LMArena leaderboard dataset
87Mistral Medium 3.5 (128B) Mistral55±2*api $3.00/M +local~42GB262kdocs.mistral.ai ↗LMArena leaderboard dataset
89Gemini 2.5 Flash Google55±2api $0.85/M1Mai.google.dev ↗LMArena leaderboard datasetGoogle DeepMind Gemini model page and model cards (lab-reported scores)ServiceNow-AI/Apriel-1.6-15b-Thinker model cardBEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (Ai2, arXiv 2601.10611v4)
90GPT-4.1 mini OpenAI54±2api $0.70/M1Mplatform.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)
91Grok 4.1 Fast (Reasoning) SpaceXAI (xAI)54±2*api $0.28/M128kazure.microsoft.com ↗LMArena leaderboard dataset
93GPT-5.4 nano OpenAI54±2*api $0.46/M400kplatform.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)
94Mistral Large 3 (675B-A41B) Mistral54±2*api $0.75/M~108GB262kdocs.mistral.ai ↗LMArena leaderboard datasetmistralai/Mistral-Small-4-119B-2603 model card
95o1 (2024-12-17) OpenAI54±2api $26.25/M200kdevelopers.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)PhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understanding (arXiv:2501.16411v2)
96Qwen3-VL 32B Thinking · 4-bit Alibaba54±2†local~24GB256kOllama › 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)
97Qwen3.5 4B · 8-bit Alibaba53±2local~6.2GB262kllama.cpp › Qwen/Qwen3.5-9B model cardChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)Qwen/Qwen3.5-4B model cardBEAR-Bench: A Bilingual Enterprise and Academic Reasoning Benchmark for Multimodal Models (arXiv:2608.17895v1)SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
99GPT-5 mini OpenAI53±2api $0.69/M400kplatform.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)VisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback? (arXiv:2608.10408v1)Qwen/Qwen3.5-27B model cardMolmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (Ai2, arXiv 2601.10611v4)
101Qwen3-VL 32B Instruct · 4-bit Alibaba52±2†local~24GB256kOllama › 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)
103Mistral Small 4 (119B-A6.5B) Mistral52±7*†api $0.26/M +local~40GB262kdocs.mistral.ai ↗mistralai/Mistral-Small-4-119B-2603 model card
106GLM-4.6V Z.ai51±2api $0.45/M +local~51GB128kdocs.z.ai ↗LMArena leaderboard datasetZ.ai GLM-V GitHub repository benchmark image (GLM-4.6V)SciDocBench: A Workflow-Centered Benchmark and Data Pipeline for Scientific Document Understanding (arXiv:2609.05141v2)
107Apriel 1.6 15B Thinker · 8-bit ServiceNow50±4*†local~21GBllama.cpp › ServiceNow-AI/Apriel-1.6-15b-Thinker model card
108GLM-4.5V Z.ai50±2api $0.90/M~69GB64kdocs.z.ai ↗LMArena leaderboard datasetZ.ai GLM-V GitHub repository benchmark image (GLM-4.6V)ServiceNow-AI/Apriel-1.6-15b-Thinker model cardChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)
110Qwen3-VL 30B-A3B Thinking · 4-bit Alibaba49±2†local~23GB256kOllama › 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)
111GPT-4o (2024-11-20) OpenAI49±4*†api $4.38/M128kdevelopers.openai.com ↗OpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)
114GPT-5 nano OpenAI47±2*api $0.14/M400kplatform.openai.com ↗LMArena leaderboard datasetQwen/Qwen3.5-4B model card
115Qwen3-VL 8B Thinking · 8-bit Alibaba47±2local~12GB256kOllama › 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)
116Llama 4 Maverick (402B-A17B) Meta47±2api $0.42/M~65GB1Mdocs.aws.amazon.com ↗LMArena leaderboard datasetApriel-1.5-15b-Thinker (arXiv:2510.01141v1), Table 4 multimodal benchmarksVMMU: A Vietnamese Multitask Multimodal Understanding and Reasoning Benchmark (arXiv:2508.13680v5)
119Qwen3-VL 30B-A3B Instruct · 4-bit Alibaba46±2†local~23GB256kOllama › 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)
120Magistral Small 1.2 · 4-bit Mistral46±5*local~17GB131kllama.cpp › Apriel-1.5-15b-Thinker (arXiv:2510.01141v1), Table 4 multimodal benchmarks
121GPT-4o (2024-05-13) OpenAI46±2api $7.50/M128kdevelopers.openai.com ↗LMArena leaderboard datasetVHELM: A Holistic Evaluation of Vision Language Models (Stanford CRFM), arXiv 2410.07112v2CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs (arXiv:2406.18521v1)Pixtral 12B (Mistral AI, arXiv 2410.07073v2)CharXiv leaderboard (princeton-nlp/CharXiv README)Microsoft Phi-3.5-vision-instruct model card: Benchmarks table
123Gemini 2.5 Flash-Lite Google45±3*api $0.17/M1Mai.google.dev ↗Google DeepMind Gemini model page and model cards (lab-reported scores)Qwen/Qwen3.5-4B model card
124Llama 4 Scout (109B-A17B) Meta45±2*api $0.15/M +local~36GB328kdeepinfra.com ↗LMArena leaderboard datasetVMMU: A Vietnamese Multitask Multimodal Understanding and Reasoning Benchmark (arXiv:2508.13680v5)
125Mistral Small 3.2 24B · 8-bit Mistral44±2local~28GB128kllama.cpp › LMArena leaderboard datasetVMMU: A Vietnamese Multitask Multimodal Understanding and Reasoning Benchmark (arXiv:2508.13680v5)mistralai/Mistral-Small-3.2-24B-Instruct-2506 model cardmistralai/Mistral-Small-4-119B-2603 model card
126Qwen3-VL 4B Thinking · 8-bit Alibaba44±2local~6.9GB256kOllama › 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)
128Qwen2.5-VL 72B Instruct · 4-bit Alibaba44±1local (slow)~48GB33kOllama › 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)LMArena leaderboard datasetQwen2.5-VL-32B-Instruct model cardChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)VMMU: A Vietnamese Multitask Multimodal Understanding and Reasoning Benchmark (arXiv:2508.13680v5)Kaleidoscope: In-language Exams for Massively Multilingual Vision Evaluation (arXiv:2504.07072v2)+2
129Qwen3-VL 8B Instruct · 8-bit Alibaba44±2local~12GB256kOllama › 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)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
131Mistral Small 3.1 24B (2503) · 8-bit Mistral44±2local~28GB128kllama.cpp › LMArena leaderboard datasetMistral AI vendor-reported benchmark scores for its own models (Hugging Face model cards)mistralai/Mistral-Small-3.2-24B-Instruct-2506 model card
132GPT-4o (2024-08-06) OpenAI43±2api $4.38/M128kdevelopers.openai.com ↗LMArena leaderboard datasetVHELM: A Holistic Evaluation of Vision Language Models (Stanford CRFM), arXiv 2410.07112v2Meridian Labs inspect_charxiv evaluation report (GitHub README)Kaleidoscope: In-language Exams for Massively Multilingual Vision Evaluation (arXiv:2504.07072v2)CapArena: Benchmarking and Analyzing Detailed Image Captioning in the LLM Era (arXiv:2503.12329v1)
133Gemma 3 27B · 4-bit Google42±2local~19GB131kOllama › LMArena leaderboard datasetGemma Team — Gemma 4 Technical ReportGoogle DeepMind Gemma 3 270M IT model cardVisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback? (arXiv:2608.10408v1)VMMU: A Vietnamese Multitask Multimodal Understanding and Reasoning Benchmark (arXiv:2508.13680v5)Gemma Team — Gemma 3 Technical Report
134Nova Premier Amazon42±5*†api $5.00/M1Maws.amazon.com ↗Amazon vendor-reported benchmark scores for its own Nova models (Amazon Science technical reports)
135Molmo 2 8B · 8-bit Ai242±2local~11GBllama.cpp › LMArena leaderboard datasetMolmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (Ai2, arXiv 2601.10611v4)
136Qwen2.5-VL 32B Instruct · 4-bit Alibaba42±2local~23GB33kOllama › LMArena leaderboard datasetQwen2.5-VL-32B-Instruct model cardApriel-1.5-15b-Thinker (arXiv:2510.01141v1), Table 4 multimodal benchmarksChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)Kaleidoscope: In-language Exams for Massively Multilingual Vision Evaluation (arXiv:2504.07072v2)VMMU: A Vietnamese Multitask Multimodal Understanding and Reasoning Benchmark (arXiv:2508.13680v5)
138Gemma 4 E4B · 8-bit Google40±4†local~9.9GB131kllama.cpp › Gemma Team — Gemma 4 Technical Report
139Pixtral Large (2411) · 2-bit Mistral40±2local (slow)~51GBllama.cpp › LMArena leaderboard datasetMistral AI vendor-reported benchmark scores for its own models (Hugging Face model cards)
140Qwen3-VL 4B Instruct · 8-bit Alibaba40±2local~6.9GB256kOllama › 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)
141Qwen3.5 2B · 8-bit Alibaba39±2local~4.9GB262kOllama › Qwen/Qwen3.5-2B model cardCohereLabs/North-Micro-Vision-Instruct model card (benchmark table)ChartArena: Benchmarking Chart Parsing across Languages, Scenarios, and Formats (arXiv:2606.01348v3)SpaceCast-Bench: Evaluating Predictive Spatial Reasoning in Vision-Language Models (arXiv:2610.12402v1)How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings (arXiv:2605.07492v2)
142GPT-4.1 nano OpenAI38±2api $0.17/M1Mdevelopers.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)
143Qwen2-VL 72B Instruct · 4-bit Alibaba38±1local (slow)~53GBllama.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)LMArena leaderboard datasetPixtral 12B (Mistral AI, arXiv 2410.07073v2)Qwen2.5-VL-32B-Instruct model cardCapArena: Benchmarking and Analyzing Detailed Image Captioning in the LLM Era (arXiv:2503.12329v1)Ai2 Molmo-72B-0924 model card: Human Preference Elo Rating column
144GPT-4 Turbo (2024-04-09) OpenAI38±2api $15.00/M128kdevelopers.openai.com ↗LMArena leaderboard datasetVHELM: A Holistic Evaluation of Vision Language Models (Stanford CRFM), arXiv 2410.07112v2CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs (arXiv:2406.18521v1)CharXiv leaderboard (princeton-nlp/CharXiv README)Phi-3 Technical Report (Microsoft, arXiv 2404.14219v3)
145GPT-4o mini (2024-07-18) OpenAI37±2api $0.26/M128kdevelopers.openai.com ↗LMArena leaderboard datasetOpenAI vendor-reported benchmark scores for its own models (openai.com announcement posts and Deployment Safety Hub system cards)VHELM: A Holistic Evaluation of Vision Language Models (Stanford CRFM), arXiv 2410.07112v2CharXiv leaderboard (princeton-nlp/CharXiv README)Microsoft Phi-3.5-vision-instruct model card: Benchmarks tableCapArena: Benchmarking and Analyzing Detailed Image Captioning in the LLM Era (arXiv:2503.12329v1)+1
148Qwen2.5-VL 7B Instruct · 8-bit Alibaba35±2local~11GBOllama › 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)
150Qwen3-VL 2B Thinking · 8-bit Alibaba34±2local~4.2GB256kOllama › 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)
155Pixtral 12B (2409) · 16-bit Mistral30±2local~28GBhow to run › LMArena leaderboard datasetPixtral 12B (Mistral AI, arXiv 2410.07073v2)VisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback? (arXiv:2608.10408v1)Ai2 Molmo-72B-0924 model card: Human Preference Elo Rating column
156Gemma 3 12B · 8-bit Google30±4local~17GB131kOllama › Google DeepMind Gemma 3 270M IT model cardVisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback? (arXiv:2608.10408v1)Gemma Team — Gemma 3 Technical Report
159Nova Pro Amazon29±2api $1.40/M300kdocs.aws.amazon.com ↗LMArena leaderboard datasetAmazon vendor-reported benchmark scores for its own Nova models (Amazon Science technical reports)
161Qwen2-VL 7B Instruct · 8-bit Alibaba29±1local~11GBllama.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)LMArena leaderboard datasetPixtral 12B (Mistral AI, arXiv 2410.07073v2)CapArena: Benchmarking and Analyzing Detailed Image Captioning in the LLM Era (arXiv:2503.12329v1)Ai2 Molmo-72B-0924 model card: Human Preference Elo Rating column
162Nova Lite Amazon28±2*api $0.10/M300kdocs.aws.amazon.com ↗LMArena leaderboard dataset
163Ministral 3 3B · 8-bit Mistral27±4local~6.9GB262kOllama › CohereLabs/North-Micro-Vision-Instruct model card (benchmark table)
164Qwen3-VL 2B Instruct · 8-bit Alibaba27±1local~4.2GB256kOllama › 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)
165Gemma 4 E2B · 8-bit Google27±3local~6.8GB131kllama.cpp › CohereLabs/North-Micro-Vision-Instruct model card (benchmark table)Gemma Team — Gemma 4 Technical Report
166Qwen2.5-VL 3B Instruct · 8-bit Alibaba27±2local~6.3GBOllama › 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)
167Qwen3.5 0.8B · 8-bit Alibaba26±2local~3.1GB262kOllama › Qwen/Qwen3.5-2B 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)
169Aya Vision 32B Cohere25±2api price unknown~71GB16kdocs.cohere.com ↗LMArena leaderboard datasetKaleidoscope: In-language Exams for Massively Multilingual Vision Evaluation (arXiv:2504.07072v2)VMMU: A Vietnamese Multitask Multimodal Understanding and Reasoning Benchmark (arXiv:2508.13680v5)
171Molmo 7B-D (0924) · 4-bit Ai224±2local (slow)~35GBhow to run › LMArena leaderboard datasetMolmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (Ai2, arXiv 2601.10611v4)Pixtral 12B (Mistral AI, arXiv 2410.07073v2)Kaleidoscope: In-language Exams for Massively Multilingual Vision Evaluation (arXiv:2504.07072v2)Ai2 Molmo-72B-0924 model card: Human Preference Elo Rating column
173Llama 3.2 11B Vision Instruct · 8-bit Meta22±2local~16GB128kOllama › LMArena leaderboard datasetPixtral 12B (Mistral AI, arXiv 2410.07073v2)
175CogVLM2 Llama3 Chat 19B · 4-bit Zhipu AI (THUDM)21±2local (slow)~43GBhow to run › LMArena leaderboard datasetCapArena: Benchmarking and Analyzing Detailed Image Captioning in the LLM Era (arXiv:2503.12329v1)Zhipu AI (THUDM) CogVLM2 model card, Benchmark table (own models)
176MiniCPM-V 2.6 · 8-bit OpenBMB21±2local~12GBOllama › LMArena leaderboard datasetCharXiv leaderboard (princeton-nlp/CharXiv README)CapArena: 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)
177InternVL2-4B · 16-bit Shanghai AI Laboratory (OpenGVLab)20±2local~10GBhow to run › LMArena leaderboard datasetOpenGVLab InternVL2 model cards (InternVL2-4B and InternVL2-26B), Evaluation tablesMicrosoft Phi-3.5-vision-instruct model card: Benchmarks tablePhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understanding (arXiv:2501.16411v2)
178LLaVA-v1.6-34B · 4-bit LLaVA19±2local~24GB4kOllama › LMArena leaderboard datasetCharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs (arXiv:2406.18521v1)CharXiv leaderboard (princeton-nlp/CharXiv README)CapArena: Benchmarking and Analyzing Detailed Image Captioning in the LLM Era (arXiv:2503.12329v1)LLaVA repository model zoo (haotian-liu/LLaVA docs/MODEL_ZOO.md)
179Phi-3.5-vision-instruct · 16-bit Microsoft16±2local~10GBhow to run › LMArena leaderboard datasetCohereLabs/North-Micro-Vision-Instruct model card (benchmark table)Microsoft Phi-3.5-vision-instruct model card: Benchmarks tableAi2 Molmo-72B-0924 model card: Human Preference Elo Rating columnPhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understanding (arXiv:2501.16411v2)
180Qwen2-VL 2B Instruct · 4-bit Alibaba14±2†local~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)
181Gemma 3 4B · 8-bit Google14±4local~7.4GB131kOllama › Google DeepMind Gemma 3 270M IT model cardVisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback? (arXiv:2608.10408v1)Gemma Team — Gemma 3 Technical ReportVMMU: A Vietnamese Multitask Multimodal Understanding and Reasoning Benchmark (arXiv:2508.13680v5)
182Phi-3-vision-128k-instruct · 16-bit Microsoft12±2local~10GBhow to run › LMArena leaderboard datasetCharXiv leaderboard (princeton-nlp/CharXiv README)Phi-3 Technical Report (Microsoft, arXiv 2404.14219v3)PhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understanding (arXiv:2501.16411v2)
—Command A+ (05-2026) Coherenot yet calibratedapi price unknown~71GB128kdocs.cohere.com ↗Cohere vendor-reported benchmark scores for its own models (cohere.com blog)
—EXAONE 4.5 33B · gguf-q4_k_m LG AI Researchnot scored yetlocal~24GB262kllama.cpp ›

+ 44 more don't fit your machine or budget · show them

+ 2 models with fewer than 3 results left off, since one or two tests say little about overall ability · show them

184 of 202 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