AA Intelligence Index v5

What does it measure?
A single report card that combines scores from many different exams. Use it when you want to compare how capable an AI is overall — coding, science, doing work tasks, knowledge — rather than any one skill. A high score usually means an AI that is good across the board rather than exceptional at one thing.
Who evaluates it, and how?
The independent evaluator Artificial Analysis runs nine evaluations itself and takes a weighted average. v5, released on October 27, 2026, gives 25% each to knowledge work (AA-Briefcase, AutomationBench-AA, GDP.pdf), coding (FrontierCode 2.0, Terminal-Bench 4.0), science (Terminal-Bench-Science, CritPt v2) and general (AA-Omniscience, HLE-Diamond). Tests with private questions or answers make up 60% of the weight, and the numbers are third-party measurements, not vendor claims.
Reading the score
0–100. The top models cluster within a few points, so a 1–2 point gap is hard to feel; look at the score band rather than the rank. It is a text-based, English-language suite, so image, speech and Korean ability are not reflected. This ranking holds v5 scores only. The test set changed, so v4 scores are not comparable; they are kept on the separate "AA Intelligence Index v4" leaderboard.

Rank

Top score64.6
Models tested54
Scoring basisHigher is better
Last updated2026.11.21
Model release date

How high is its intelligence index across knowledge work, coding and science?

AA Intelligence Index v5

  1. 🥇AnthropicClaude Fable 5.2Reasoning effort: Max64.6150384.9
  2. 🥈OpenAIGPT-6.2 AstraReasoning effort: Max62.11468—
  3. 🥉GoogleGemini 4 NeonReasoning effort: Extra High61.41513—
  4. #4GoogleGemini 4 CarbonReasoning effort: High60.01527—
  5. #5AnthropicClaude Opus 5.5Reasoning effort: Max59.1150683.4
  6. #6OpenAIGPT-6.2 SolReasoning effort: Max57.31465—
  7. #7OpenAIGPT-6 AstraReasoning effort: Max56.9147584.5
  8. #8OpenAIGPT-6.1 SolReasoning effort: Max54.6148381.4
  9. #9AnthropicClaude Sonnet 5.5Reasoning effort: Max54.5147377.0
  10. #10GoogleGemini 4 HydrogenReasoning effort: Extra High53.91490—
  11. #11OpenAIGPT-6.1 TerraReasoning effort: Max51.3147478.8
  12. #12GoogleGemini 4 ArgonReasoning effort: High50.4152077.7
  13. #13AnthropicClaude Fable 5.1Reasoning effort: Extra High50.0150384.0
  14. #14Moonshot AIKimi K3.1 Pro46.5148268.7
  15. #15AnthropicClaude Opus 5Reasoning effort: Max44.21489—
  16. #15OpenAIGPT-6 SolReasoning effort: Max44.21458—
  17. #17MetaMuse WonderReasoning effort: Max43.61511—
  18. #18AlibabaQwen4 Max42.9——
  19. #19MetaMuse Spark 1.3Reasoning effort: Max36.6149868.4
  20. #20AnthropicClaude Haiku 5.5Reasoning effort: Max36.5142169.0
  21. #21MetaMuse EmberReasoning effort: Max36.0——
  22. #22SpaceXSIGrok 4.7Reasoning effort: Extra High35.5144367.9
  23. #23TencentHy434.6——
  24. #24DeepSeekDeepSeek V4.1 Flash 1017Reasoning effort: Max32.6—70.3
  25. #25OpenAIGPT-5.6 TerraReasoning effort: Max32.3146867.2
  26. #26Z.aiGLM 5.3Reasoning effort: Max31.71482—
  27. #27SpaceXSIGrok 4.6Reasoning effort: Extra High31.0145763.5
  28. #28GoogleGemini 3.8 FlashReasoning effort: High30.8149756.0
  29. #29InklingInkling 2 Small30.61412—
  30. #30XiaomiMiMo-V2.6-Pro29.9148166.1
  31. #31MiniMaxMiniMax M3-1 Flash29.1—63.7
  32. #32AlibabaQwen3.8 Flash28.6——
  33. #33StepfunStep-528.51449—
  34. #34OpenAIGPT-6 LunaReasoning effort: Max27.9144360.4
  35. #35Z.aiGLM 5.3 Flash27.8147560.7
  36. #36XiaomiMiMo-V2.6-Flash26.71448—
  37. #37Mistral AIMistral Large 425.3142340.9
  38. #38OpenAIGPT-5.6 LunaReasoning effort: Max23.81453—
  39. #39DeepSeekDeepSeek V4 ProReasoning effort: Max23.11458—
  40. #40MiniMaxMiniMax M318.91434—
  41. #41Moonshot AIKimi K3Reasoning effort: Low17.4148865.0
  42. #42UpstageSolar Pro 414.21388—
  43. #43NVIDIANemotron 3 UltraReasoning effort: Thinking11.81427—
  44. #44GoogleGemini 3.5 Flash-Lite11.71455—
  45. #45SK TelecomA.X K29.0——
  46. #46LG AI ResearchK-EXAONE 2.07.6——
  47. #47AnthropicClaude Haiku 4.5Reasoning effort: Thinking7.41414—
  48. #48GoogleGemini 3.1 Flash Lite6.21433—
  49. #49Liquid AILFM3-1.7B6.0——
  50. #50Mistral AIMistral Medium 3.54.61426—
  51. #51GoogleGemma 4 31BReasoning effort: Thinking4.4145232.4
  52. #52Mistral AIMistral Small 4Reasoning effort: Thinking4.3——
  53. #53OpenAIGPT OSS 120BReasoning effort: High3.81352—
  54. #54CohereNorth Mini Code3.4——
Half of models ≤ 30.9

39 of the 64 AIs released in the last 60 days have no score here yet · Microsoft-Decision-1.1, d1.2, Hy4 flash and 36 more

This score comes from a fixed, predefined evaluation, so it may differ from what you get with your own question.

Scores labelled 'Self-reported' were published by the AI maker itself, not measured by a third party.