All rankings / Vision
Vision · style controlled
Human preference ratings for the overall category of this Arena task. Scores from different boards are not interchangeable.
167 entries · Arena score · Higher is better
Source publication: 2026-10-19. Retrieved 2026-11-05.
Source: Arena (lmarena-ai) · CC-BY-4.0.
Ranks use confidence bounds. Several models can share a rank; the rows within a shared rank are displayed by score.
| Significance rank | Model / evaluated variant | Arena score | 95% confidence interval | Votes |
|---|---|---|---|---|
| 1 | Claude Fable 5 (High)anthropic | 1,308.71 | 1,301.66 – 1,315.77 | 14,714 |
| 1 | Qwen3.8-Maxalibaba | 1,300 | 1,292.54 – 1,307.46 | 9,957 |
| 1 | Claude Opus 4.6 (High)anthropic | 1,299.07 | 1,292.65 – 1,305.5 | 22,286 |
| 1 | Claude Opus 4.7anthropic | 1,298.72 | 1,292.15 – 1,305.3 | 23,139 |
| 1 | Claude Opus 4.7 (High)anthropic | 1,297.72 | 1,291.16 – 1,304.28 | 22,731 |
| 1 | Gemini 3.7 Flash (High)google | 1,295.17 | 1,285.94 – 1,304.4 | 5,369 |
| 1 | Muse Sparkmeta | 1,294.09 | 1,284.98 – 1,303.19 | 5,865 |
| 2 | Claude Opus 4.6anthropic | 1,293.71 | 1,287.41 – 1,300.01 | 26,841 |
| 1 | GPT-6.1 Sol (Max)openai | 1,293.53 | 1,280.59 – 1,306.47 | 2,330 |
| 1 | Claude Opus 5.5 (High)anthropic | 1,292.96 | 1,281.21 – 1,304.71 | 3,035 |
| 1 | Muse Spark 1.2 (xHigh)meta | 1,292.67 | 1,278.33 – 1,307.01 | 1,948 |
| 2 | Gemini 3.8 Flash (High)google | 1,291.04 | 1,281.23 – 1,300.84 | 4,645 |
| 2 | Muse Spark 1.3 (Max)meta | 1,289.59 | 1,280.38 – 1,298.8 | 5,392 |
| 2 | Gemini 3 Progoogle | 1,289.3 | 1,281.9 – 1,296.69 | 13,623 |
| 2 | Claude Opus 5 (High)anthropic | 1,287.86 | 1,281.2 – 1,294.51 | 17,743 |
| 2 | Muse Spark 1.3 (xHigh)meta | 1,287.63 | 1,278.08 – 1,297.18 | 4,897 |
| 2 | Claude Fable 5.1 (Max)anthropic | 1,287.59 | 1,278.26 – 1,296.91 | 5,026 |
| 2 | GPT-5.6 Sol (xHigh)openai | 1,286.16 | 1,278.83 – 1,293.49 | 11,439 |
| 5 | GPT-5.5openai | 1,286.05 | 1,280.26 – 1,291.84 | 28,781 |
| 4 | Claude Opus 4.8 (High)anthropic | 1,285.74 | 1,279.26 – 1,292.22 | 18,303 |
| 2 | GPT-6 Astra (Max)openai | 1,284.82 | 1,274.49 – 1,295.15 | 4,012 |
| 6 | GPT-5.4 (High)openai | 1,284.14 | 1,278.37 – 1,289.9 | 30,747 |
| 6 | Gemini 3.5 Flash (Medium)google | 1,282.39 | 1,275.58 – 1,289.19 | 14,061 |
| 6 | Gemini 3.5 Flash (High)google | 1,282.22 | 1,275.36 – 1,289.08 | 13,862 |
| 6 | GPT-5.5 (High)openai | 1,282.16 | 1,276.28 – 1,288.03 | 27,067 |
| 6 | Gemini 3.6 Flash (High)google | 1,281.98 | 1,274.09 – 1,289.88 | 8,979 |
| 7 | Muse Spark 1.1meta | 1,280.01 | 1,273.05 – 1,286.96 | 12,804 |
| 9 | Gemini 3.1 Pro Previewgoogle | 1,279.55 | 1,274.42 – 1,284.68 | 47,114 |
| 7 | Grok 4.5xai | 1,278.98 | 1,271.89 – 1,286.06 | 12,599 |
| 8 | Claude Opus 4.8anthropic | 1,278.86 | 1,272.35 – 1,285.37 | 18,800 |
| 8 | GPT-5.4openai | 1,278.64 | 1,272.13 – 1,285.15 | 22,804 |
| 9 | GPT-5.2 Chat (20260210)openai | 1,277.7 | 1,271.17 – 1,284.22 | 16,465 |
| 9 | GPT-5.5 Instantopenai | 1,276.06 | 1,267.32 – 1,284.8 | 7,590 |
| 11 | Claude Sonnet 4.6anthropic | 1,275.09 | 1,268.95 – 1,281.22 | 27,408 |
| 22 | Gemini 3 Flashgoogle | 1,272.54 | 1,267.34 – 1,277.74 | 39,131 |
| 22 | GLM-5.3-Flashzai | 1,271.15 | 1,264.75 – 1,277.56 | 16,254 |
| 25 | GPT-5.6 Terra (xHigh)openai | 1,268.04 | 1,260.77 – 1,275.31 | 11,842 |
| 25 | Gemini 3.5 Flash-Litegoogle | 1,266.95 | 1,259.08 – 1,274.81 | 9,238 |
| 22 | GPT-6 Sol (Max)openai | 1,266.05 | 1,254.69 – 1,277.4 | 3,221 |
| 22 | Claude Sonnet 5.5 (xHigh)anthropic | 1,264.92 | 1,252.36 – 1,277.47 | 2,564 |
| 32 | Kimi K2.6moonshot | 1,264.57 | 1,257.76 – 1,271.38 | 16,368 |
| 33 | Qwen3.7-Plusalibaba | 1,263.31 | 1,256.61 – 1,270.01 | 14,212 |
| 33 | Grok 4.6 (High)xai | 1,262.14 | 1,253.82 – 1,270.47 | 7,261 |
| 34 | Claude Sonnet 5 (High)anthropic | 1,261.86 | 1,255.03 – 1,268.7 | 14,034 |
| 34 | Gemma 4 31Bgoogle | 1,261.85 | 1,256.14 – 1,267.55 | 40,219 |
| 32 | DeepSeek-V4.1-Flash-20260910 (Max)deepseek | 1,261.36 | 1,251.08 – 1,271.64 | 4,100 |
| 36 | Gemini 3 Flash (Minimal)google | 1,259.97 | 1,254.65 – 1,265.28 | 37,173 |
| 37 | GPT-5.6 Luna (xHigh)openai | 1,256.59 | 1,249.15 – 1,264.03 | 11,055 |
| 37 | Dola Seed 2.0 Probytedance | 1,256.47 | 1,249.21 – 1,263.74 | 10,972 |
| 37 | Grok 4.20 Beta (Reasoning, 0309)xai | 1,255.19 | 1,249.11 – 1,261.28 | 26,796 |
| 39 | Kimi K2.5 (Thinking)moonshot | 1,252.36 | 1,246.68 – 1,258.04 | 28,249 |
| 32 | Mistral Large 4mistral | 1,251.94 | 1,232.61 – 1,271.28 | 1,008 |
| 39 | GPT-5.1 (High)openai | 1,251.2 | 1,243.37 – 1,259.03 | 9,848 |
| 40 | GPT-5.4 mini (High)openai | 1,250.44 | 1,244.22 – 1,256.66 | 25,734 |
| 41 | Grok 4.20 Multi-Agent Beta (0309)xai | 1,250.22 | 1,244.1 – 1,256.33 | 24,532 |
| 38 | Grok 4.7 (xHigh)xai | 1,249.77 | 1,239.15 – 1,260.39 | 3,736 |
| 47 | Gemini 2.5 Progoogle | 1,247.64 | 1,243.09 – 1,252.19 | 88,755 |
| 37 | Step 5 Previewstepfun | 1,247.18 | 1,232.02 – 1,262.33 | 1,649 |
| 47 | Qwen3.5-397B-A17Balibaba | 1,245.81 | 1,240.51 – 1,251.12 | 32,657 |
| 38 | MiMo-V2.6-Proxiaomi | 1,244.56 | 1,229.79 – 1,259.34 | 1,754 |
| 42 | MiMo-V2.6-Flashxiaomi | 1,244.53 | 1,233.65 – 1,255.42 | 3,618 |
| 51 | GPT-5.2 (High)openai | 1,242.8 | 1,236.71 – 1,248.88 | 19,022 |
| 48 | Qwen3.8-27Balibaba | 1,242.43 | 1,234.22 – 1,250.65 | 7,614 |
| 51 | Grok 4.3xai | 1,242.27 | 1,236.42 – 1,248.11 | 27,901 |
| 51 | Gemma 4 26B A4Bgoogle | 1,241.74 | 1,235.28 – 1,248.2 | 23,922 |
| 47 | GPT-6 Luna (Max)openai | 1,240.36 | 1,229.47 – 1,251.25 | 3,662 |
| 51 | GPT-5.1openai | 1,239.84 | 1,232.29 – 1,247.38 | 10,801 |
| 54 | ChatGPT-4o Latest (20250326)openai | 1,238.59 | 1,233.09 – 1,244.09 | 23,400 |
| 51 | Kimi K2.5 (Instant)moonshot | 1,236.47 | 1,225.76 – 1,247.19 | 4,111 |
| 57 | Gemini 3.1 Flash-Lite Previewgoogle | 1,234.63 | 1,228.88 – 1,240.38 | 27,657 |
| 57 | MiMo-V2.5xiaomi | 1,233.76 | 1,227.92 – 1,239.6 | 29,381 |
| 57 | MiniMax-M3minimax | 1,233.29 | 1,226.9 – 1,239.68 | 19,844 |
| 60 | GLM-5V-Turbozai | 1,230.04 | 1,224.56 – 1,235.52 | 42,757 |
| 57 | Gemini 2.5 Flash Preview (09-2025)google | 1,229.8 | 1,219.96 – 1,239.64 | 4,792 |
| 62 | GPT-5.2openai | 1,228.15 | 1,222.15 – 1,234.14 | 20,100 |
| 64 | Qwen3.5-122B-A10Balibaba | 1,226.17 | 1,219.4 – 1,232.95 | 13,236 |
| 64 | GPT-5 Chatopenai | 1,225.84 | 1,218.78 – 1,232.9 | 39,894 |
| 59 | GPT-4.5 Preview (2025-02-27)openai | 1,224.99 | 1,213.29 – 1,236.68 | 2,925 |
| 67 | ERNIE-5.0-Preview (1220)baidu | 1,219.42 | 1,208.25 – 1,230.6 | 3,349 |
| 74 | Qwen3.5-27Balibaba | 1,217.16 | 1,211.65 – 1,222.68 | 30,272 |
| 74 | MiMo-V2-Omnixiaomi | 1,215.89 | 1,208.53 – 1,223.26 | 10,650 |
| 74 | o3 (2025-04-16)openai | 1,215.85 | 1,209.27 – 1,222.43 | 45,567 |
| 75 | Gemini 2.5 Flashgoogle | 1,215.46 | 1,210.87 – 1,220.04 | 60,087 |
| 75 | Qwen3-VL-235B-A22B-Instructalibaba | 1,215.1 | 1,208.11 – 1,222.1 | 12,724 |
| 75 | GPT-4.1 (2025-04-14)openai | 1,213.87 | 1,207.18 – 1,220.57 | 40,998 |
| 78 | GPT-5 (High)openai | 1,210.72 | 1,203.66 – 1,217.78 | 34,445 |
| 78 | Inkling-Smallthinky | 1,208.75 | 1,200.57 – 1,216.93 | 8,074 |
| 74 | Claude Opus 4 (Thinking 16K, 20250514)anthropic | 1,207.5 | 1,191.44 – 1,223.56 | 1,290 |
| 74 | Claude Sonnet 4 (Thinking 32K, 20250514)anthropic | 1,206.55 | 1,189.62 – 1,223.48 | 1,144 |
| 78 | NVIDIA-Nemotron-3.5-Super-VL-120B-A12B (NVFP4)nvidia | 1,204.01 | 1,191.15 – 1,216.86 | 2,509 |
| 85 | o4-mini (2025-04-16)openai | 1,200.95 | 1,193.82 – 1,208.09 | 41,316 |
| 83 | GPT-4.1 mini (2025-04-14)openai | 1,200.74 | 1,193.22 – 1,208.27 | 40,201 |
| 82 | Mistral Medium 3.5mistral | 1,199.59 | 1,190.02 – 1,209.17 | 5,375 |
| 86 | Mistral Large 3mistral | 1,199.09 | 1,191.3 – 1,206.89 | 9,254 |
| 86 | GPT-5.4 nano (High)openai | 1,198.84 | 1,192.5 – 1,205.18 | 25,624 |
| 87 | Grok 4.1 Fast (Reasoning)xai | 1,196.34 | 1,189.42 – 1,203.25 | 13,368 |
| 84 | Claude 3.7 Sonnet (Thinking 32K, 20250219)anthropic | 1,192.66 | 1,177.15 – 1,208.17 | 1,436 |
| 87 | o1 (2024-12-17)openai | 1,192.66 | 1,181.89 – 1,203.42 | 3,694 |
| 87 | Claude Opus 4 (20250514)anthropic | 1,188.94 | 1,176.03 – 1,201.85 | 2,172 |
| 87 | Qwen3-VL-235B-A22B-Thinkingalibaba | 1,188.78 | 1,176.44 – 1,201.12 | 2,387 |
| 88 | Gemini 2.5 Flash-Lite Preview (Thinking, 06-17)google | 1,188.75 | 1,181.28 – 1,196.23 | 35,914 |
| 87 | Claude Sonnet 4 (20250514)anthropic | 1,188.55 | 1,174.51 – 1,202.59 | 1,806 |
| 88 | Qwen-VL-Max (2025-08-13)alibaba | 1,185.22 | 1,173.13 – 1,197.32 | 3,170 |
| 91 | GPT-5 mini (High)openai | 1,184.32 | 1,176.22 – 1,192.41 | 28,943 |
| 91 | Grok 4 (0709)xai | 1,184.2 | 1,176.84 – 1,191.55 | 30,173 |
| 97 | Gemini 1.5 Pro (002)google | 1,179.47 | 1,170.39 – 1,188.56 | 8,902 |
| 97 | Claude 3.7 Sonnet (20250219)anthropic | 1,175.11 | 1,165.73 – 1,184.49 | 4,517 |
| 97 | Gemini 2.5 Flash-Lite Preview (Non-Thinking, 09-2025)google | 1,174.57 | 1,164.97 – 1,184.18 | 4,862 |
| 99 | Gemini 2.0 Flash (001)google | 1,172.04 | 1,164.81 – 1,179.28 | 10,135 |
| 106 | GPT-4o (2024-05-13)openai | 1,161.77 | 1,153 – 1,170.53 | 23,273 |
| 104 | GLM-4.6Vzai | 1,161.4 | 1,147.64 – 1,175.17 | 2,455 |
| 107 | Claude 3.5 Sonnet (20241022)anthropic | 1,160.75 | 1,152.81 – 1,168.69 | 10,374 |
| 107 | Gemma 3 27B ITgoogle | 1,159.9 | 1,151.79 – 1,168.01 | 16,982 |
| 108 | Mistral Medium 3.1 (2508)mistral | 1,159.52 | 1,153.3 – 1,165.73 | 41,611 |
| 107 | Step-1o Turbo (202506)stepfun | 1,155.93 | 1,141.59 – 1,170.28 | 1,895 |
| 110 | Mistral Medium 3 (2505)mistral | 1,155.41 | 1,147.11 – 1,163.7 | 10,793 |
| 107 | Hunyuan Vision 1.5 Thinkingtencent | 1,154.86 | 1,142.62 – 1,167.11 | 2,424 |
| 109 | GLM-4.5Vzai | 1,153.08 | 1,141.32 – 1,164.84 | 3,243 |
| 107 | Hunyuan Large Visiontencent | 1,149.77 | 1,133.28 – 1,166.27 | 1,274 |
| 110 | GPT-5 nano (High)openai | 1,147.43 | 1,136.09 – 1,158.78 | 3,928 |
| 110 | Llama 4 Maverick 17B 128E Instructmeta | 1,145.45 | 1,136.33 – 1,154.57 | 6,923 |
| 110 | Claude 3.5 Sonnet (20240620)anthropic | 1,145.23 | 1,135.15 – 1,155.32 | 21,624 |
| 110 | Step-3stepfun | 1,144.88 | 1,132.95 – 1,156.81 | 3,193 |
| 114 | Gemini 1.5 Flash (002)google | 1,141.69 | 1,131.75 – 1,151.62 | 7,241 |
| 114 | Mistral Small 3.2 (2506)mistral | 1,139.86 | 1,130.76 – 1,148.96 | 10,843 |
| 116 | Gemini 2.0 Flash-Lite Preview (02-05)google | 1,135.82 | 1,125.36 – 1,146.27 | 3,991 |
| 119 | Llama 4 Scout 17B 16E Instructmeta | 1,127.52 | 1,118.02 – 1,137.02 | 6,446 |
| 116 | Claude 3.5 Haiku (20241022)anthropic | 1,127.41 | 1,111.36 – 1,143.46 | 1,398 |
| 121 | Mistral Small 3.1 24B Instruct (2503)mistral | 1,127.35 | 1,118.85 – 1,135.86 | 28,490 |
| 119 | Step-1o Vision 32K (High Res)stepfun | 1,124.16 | 1,111.51 – 1,136.81 | 2,833 |
| 124 | Qwen2.5-VL-72B-Instructalibaba | 1,121.11 | 1,110.29 – 1,131.94 | 3,768 |
| 125 | Gemini 1.5 Pro (001)google | 1,119.53 | 1,107.9 – 1,131.17 | 16,734 |
| 122 | Qwen2.5-VL-32B-Instructalibaba | 1,118.8 | 1,103.32 – 1,134.29 | 1,490 |
| 126 | GPT-4o (2024-08-06)openai | 1,117.92 | 1,105.63 – 1,130.2 | 3,376 |
| 125 | Molmo 2 8Ballenai | 1,111.95 | 1,092.22 – 1,131.68 | 1,190 |
| 127 | GPT-4 Turbo (2024-04-09)openai | 1,111.94 | 1,100.35 – 1,123.52 | 13,391 |
| 134 | GPT-4o mini (2024-07-18)openai | 1,096.87 | 1,088.57 – 1,105.18 | 17,341 |
| 134 | Pixtral Large (2411)mistral | 1,095.37 | 1,086.03 – 1,104.7 | 5,423 |
| 133 | GPT-4.1 nano (2025-04-14)openai | 1,088.55 | 1,070.26 – 1,106.84 | 1,211 |
| 136 | Qwen2-VL-72Balibaba | 1,084.58 | 1,074.4 – 1,094.75 | 5,937 |
| 135 | Qwen-VL-Max (1119)alibaba | 1,084.55 | 1,068.5 – 1,100.59 | 1,422 |
| 139 | Gemini 1.5 Flash-8B (001)google | 1,071.27 | 1,060.66 – 1,081.88 | 6,243 |
| 139 | Step-1V 32Kstepfun | 1,064.15 | 1,047.4 – 1,080.91 | 1,534 |
| 140 | Claude 3 Opus (20240229)anthropic | 1,061.51 | 1,051.11 – 1,071.92 | 15,565 |
| 140 | Gemini 1.5 Flash (001)google | 1,059.13 | 1,047.25 – 1,071.02 | 13,260 |
| 143 | Molmo 72B (0924)allenai | 1,044.84 | 1,031.74 – 1,057.93 | 3,048 |
| 142 | Hunyuan Standard Vision (2024-12-31)tencent | 1,041.5 | 1,020.5 – 1,062.51 | 809 |
| 146 | Llama 3.2 90B Vision Instructmeta | 1,032.21 | 1,022.96 – 1,041.45 | 8,682 |
| 146 | Qwen2-VL-7B-Instructalibaba | 1,030.64 | 1,020.05 – 1,041.23 | 5,766 |
| 146 | Pixtral 12B (2409)mistral | 1,025.19 | 1,015.63 – 1,034.75 | 7,511 |
| 146 | InternVL2 26BOrganization not supplied | 1,024.85 | 1,011.98 – 1,037.72 | 5,148 |
| 146 | Amazon Nova Lite 1.0amazon | 1,018.73 | 1,003.69 – 1,033.78 | 1,854 |
| 146 | Amazon Nova Pro 1.0amazon | 1,018.47 | 1,004.64 – 1,032.3 | 2,335 |
| 147 | Claude 3 Sonnet (20240229)anthropic | 1,016.35 | 1,004.95 – 1,027.74 | 12,314 |
| 148 | Yi VisionOrganization not supplied | 1,002.39 | 984.27 – 1,020.51 | 1,219 |
| 148 | Aya Vision 32Bcohere | 1,000.36 | 978.36 – 1,022.37 | 847 |
| 151 | Claude 3 Haiku (20240307)anthropic | 999.71 | 987.26 – 1,012.17 | 13,380 |
| 152 | Molmo 7B-D (0924)allenai | 993.72 | 979.93 – 1,007.5 | 2,815 |
| 154 | Llama 3.2 11B Vision Instructmeta | 992.75 | 981.43 – 1,004.07 | 4,817 |
| 152 | nvila-internal-15b-v1nvidia | 985.16 | 965.04 – 1,005.28 | 1,077 |
| 155 | LLaVA-OneVision Qwen2 72B OVOrganization not supplied | 977.93 | 959.74 – 996.13 | 1,321 |
| 160 | LLaVA 1.6 34BOrganization not supplied | 964.35 | 952.1 – 976.6 | 4,531 |
| 159 | CogVLM2 Llama3 Chat 19Bzai | 963.65 | 948.44 – 978.87 | 1,991 |
| 160 | MiniCPM-V 2.6Organization not supplied | 962.37 | 946.53 – 978.21 | 1,987 |
| 160 | InternVL2 4BOrganization not supplied | 958.37 | 945.73 – 971 | 3,703 |
| 166 | Phi-3.5 Vision Instructmicrosoft | 919 | 903.18 – 934.83 | 2,592 |
| 167 | Phi-3 Vision 128K Instructmicrosoft | 880.54 | 861.9 – 899.18 | 1,401 |
What this ranking can tell you
- The population is the models in this source board, including variants and models without an available API price.
- Ranks use the published confidence bounds: one plus the number of models whose lower bound exceeds this model’s upper bound. Shared ranks are not proof of equal performance.
Check the source
The original evidence is retained byte for byte. The downloadable board includes the source URLs, revision and SHA-256 hashes.
Download this board and provenance →