Skip to content
Model Index
  • Overview
  • Models
  • Compare
  • Benchmarks
  • Finder
LabsMethodology
  1. Benchmarks
  2. Vision

Arena · Vision

Human preferenceArena · updated Oct 2, 2026

What it measures

Blind head-to-head votes on answers to prompts that include images.

How to read it

Human preference for image understanding. Several of the newest models are not listed yet, so coverage is thinner than the text board.

Open at Arena
Best
1301
Qwen3.8-Max
Median
1268
GPT-5.6 Terra

Results

19 of 41 tracked models have a result. This scale has no meaningful zero, so bars run from the lowest to the highest tracked score.

Filter in explorer
  1. 1
    Qwen3.8-Max

    Alibaba · Max effort · 10,040 votes

    1301 ±7
  2. 2
    GPT-6.1 Sol

    OpenAI · Max effort · 1,408 votes

    1291 ±16
  3. 3
    Gemini 3.8 Flash

    Google DeepMind · High effort · 3,086 votes

    1290 ±12
  4. 3
    Muse Spark 1.3

    Meta · Max effort · 3,843 votes

    1290 ±10
  5. 5
    Claude Fable 5.1

    Anthropic · Max effort · 4,312 votes

    1288 ±10
  6. 6
    GPT-6 Astra

    OpenAI · Max effort · 3,175 votes

    1284 ±11
  7. 7
    Gemini 3.1 Pro

    Google DeepMind · 45,558 votes

    1279 ±5
  8. 8
    GLM-5.3-FlashOpen

    Z.ai · 5,479 votes

    1278 ±9
  9. 9
    Claude Sonnet 5.5

    Anthropic · xHigh effort · 1,608 votes

    1268 ±15
  10. 9
    GPT-5.6 Terra

    OpenAI · xHigh effort · 10,292 votes

    1268 ±8
  11. 11
    Gemini 3.5 Flash-Lite

    Google DeepMind · 7,579 votes

    1264 ±8
  12. 12
    Gemma 4 31BOpen

    Google DeepMind · 39,111 votes

    1262 ±6
  13. 13
    Step 5 Preview

    StepFun · 604 votes

    1260 ±24
  14. 14
    MiMo-V2.6-FlashOpen

    Xiaomi · 2,275 votes

    1247 ±13
  15. 15
    MiMo-V2.6-ProOpen

    Xiaomi · 1,787 votes

    1245 ±15
  16. 16
    Qwen3.8-27BOpen

    Alibaba · 6,422 votes

    1243 ±9
  17. 17
    MiniMax-M3Open

    MiniMax · 18,308 votes

    1237 ±7
  18. 18
    TM
    Inkling-SmallOpen

    Thinking Machines Lab · 6,904 votes

    1206 ±9
  19. 19
    Mistral Medium 3.5Open

    Mistral AI · 5,379 votes

    1199 ±10
22 tracked models without a result

Claude Haiku 5.5, Claude Opus 5.5, Command A+, DeepSeek V4 Pro, DeepSeek V4.1 Flash, Gemini 4 Argon, GLM-5.3, GPT-6 Luna, gpt-oss-120b, Grok 4.7, Hy3, Hy4 preview, Inkling, Kimi K2.7 Code, Kimi K3, MAI-Thinking-1, Mistral Large 4, Mistral Small 4, Muse Glimmer 30B, Nemotron 3 Ultra, Nova 2 Lite, Qwen3.8-Flash.

Missing means the evaluator has not published a run, not that the model would score poorly.

Model Index

An independent snapshot of frontier SI models and how they perform on third-party benchmarks.

Data as of Oct 8, 2026

Explore

  • All models
  • Compare
  • Benchmarks
  • Model finder
  • Labs
  • Methodology & sources

Sources

Benchmark results come from Artificial Analysis, Arena, Vals AI and ARC Prize. Specifications and prices come from each lab's own documentation.

Not affiliated with any lab or evaluator. Names and logos are trademarks of their respective owners and are used only to identify them. Scores change often — check the linked source before making a decision that depends on one.

  • Overview
  • Models
  • Compare
  • Benchmarks
  • Finder
  • Overview
  • Models
  • Compare
  • Benchmarks
  • Finder
  • Overview
  • Models
  • Compare
  • Benchmarks
  • Finder