Skip to content
Model Index
  • Overview
  • Models
  • Compare
  • Benchmarks
  • Finder
LabsMethodology
  1. Benchmarks
  2. Knowledge & reliability

AA-LCR

Independent testArtificial Analysis · updated Oct 8, 2026

What it measures

Reasoning questions (v1.1) whose answers require combining information spread across very long document sets.

How to read it

A large context window only matters if the model can use it. Scores cluster in the high 70s to high 80s.

Open at Artificial Analysis
Best
88.7%
Kimi K3
Median
80.3%
DeepSeek V4 Pro

Results

39 of 41 tracked models have a result.

Filter in explorer
  1. 1
    Kimi K3Open

    Moonshot AI · Max effort

    88.7%
  2. 2
    Step 5 Preview

    StepFun

    88.3%
  3. 3
    MiMo-V2.6-ProOpen

    Xiaomi

    86.3%
  4. 4
    Claude Fable 5.1

    Anthropic · Max effort

    85.3%
  5. 5
    Claude Opus 5.5

    Anthropic · Max effort

    84.7%
  6. 6
    DeepSeek V4.1 FlashOpen

    DeepSeek · Max effort

    84.0%
  7. 7
    GPT-6 Luna

    OpenAI · Max effort

    83.3%
  8. 7
    Muse Glimmer 30BOpen

    Meta · High effort

    83.3%
  9. 9
    GPT-5.6 Terra

    OpenAI · Max effort

    83.0%
  10. 9
    GPT-6.1 Sol

    OpenAI · Max effort

    83.0%
  11. 9
    MiniMax-M3Open

    MiniMax

    83.0%
  12. 9
    Muse Spark 1.3

    Meta · Max effort

    83.0%
  13. 13
    Claude Haiku 5.5

    Anthropic · Max effort

    82.7%
  14. 13
    Claude Sonnet 5.5

    Anthropic · Max effort

    82.7%
  15. 15
    Gemini 3.1 Pro

    Google DeepMind

    82.0%
  16. 15
    Qwen3.8-27BOpen

    Alibaba · xHigh effort

    82.0%
  17. 17
    Gemini 3.8 Flash

    Google DeepMind · High effort

    81.3%
  18. 17
    Mistral Large 4

    Mistral AI

    81.3%
  19. 19
    GPT-6 Astra

    OpenAI · Max effort

    80.7%
  20. 20
    DeepSeek V4 ProOpen

    DeepSeek · Max effort

    80.3%
  21. 20
    Qwen3.8-Max

    Alibaba

    80.3%
  22. 22
    GLM-5.3-FlashOpen

    Z.ai

    80.0%
  23. 23
    Gemini 4 ArgonLimited

    Google DeepMind · High effort

    79.7%
  24. 23
    GLM-5.3Open

    Z.ai · Max effort

    79.7%
  25. 23
    Qwen3.8-FlashOpen

    Alibaba

    79.7%
  26. 26
    Kimi K2.7 CodeOpen

    Moonshot AI

    79.3%
  27. 26
    Nemotron 3 UltraOpen

    NVIDIA · Reasoning effort

    79.3%
  28. 28
    Hy3Open

    Tencent

    79.0%
  29. 29
    TM
    InklingOpen

    Thinking Machines Lab · xHigh effort

    77.3%
  30. 30
    Grok 4.7

    SpaceXAI · xHigh effort

    76.7%
  31. 31
    Gemini 3.5 Flash-Lite

    Google DeepMind

    76.0%
  32. 32
    TM
    Inkling-SmallOpen

    Thinking Machines Lab

    75.7%
  33. 33
    MiMo-V2.6-FlashOpen

    Xiaomi

    74.3%
  34. 34
    Gemma 4 31BOpen

    Google DeepMind · Reasoning effort

    69.7%
  35. 35
    Mistral Medium 3.5Open

    Mistral AI

    69.3%
  36. 36
    Nova 2 Lite

    Amazon · High effort

    60.3%
  37. 37
    Command A+Open

    Cohere

    52.7%
  38. 38
    gpt-oss-120bOpen

    OpenAI · High effort

    52.0%
  39. 39
    Mistral Small 4Open

    Mistral AI · Reasoning effort

    49.7%
2 tracked models without a result

Hy4 preview, MAI-Thinking-1.

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