Benchmark Almanac

CompassRank

Claude Opus 5 83.3 leads since July 2026
Not in model cards

OpenCompass ranks open and API language models on its own evaluation sets and publishes monthly rankings. Model makers submit a repository link or API endpoint to be tested.

The OpenCompass README calls it the home of detailed leaderboards for open-source and API models. The first OpenCompass LLM leaderboard space dates from September 8th, 2023.

Released
September 2023
Kind
Composite indices
Level
Expert
Scored
Checked automatically
Answers
Short answer
Tasks
Kept private
Strongest model tested
Claude Opus 5 #10 on the capability index
Last model added
August 2026
Scores refreshed
November 12th, 2026
15
Entries on its own leaderboard, 14 models we track

Claude Opus 5 leads with 83.3 among models released by August 2026

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.

The newest model with a score here came out in August 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

15 entries. Scroll sideways for the rest.

  1. 83.3 Claude Opus 5
  2. 80.8 GPT-5.6 Sol
  3. 80.6 GPT-5.5
  4. 80.6 Qwen 3.8 Max
  5. 80.5 Doubao-Seed-2.1-Pro-260628 (high)
  6. 80.5 Gemini 3.7 Flash
  7. 79.1 Kimi K3
  8. 78.0 Grok 4.6
  9. 77.5 DeepSeek-V4-Pro
  10. 75.7 Kimi K2.6
  11. 75.1 Claude Opus 4.8
  12. 74.4 GLM-5.3
  13. 73.7 DeepSeek-V4-Flash
  14. 73.6 Tencent Hy3 preview
  15. 69.0 MiniMax-M3
  1. 1 Claude Opus 5 Source edition 26-07 83.3
  2. 2 GPT-5.6 Sol Source edition 26-07 80.8
  3. 3 GPT-5.5 Source edition 26-07 80.6
  4. 4 Qwen 3.8 Max Source edition 26-07 80.6
  5. 5 Doubao-Seed-2.1-Pro-260628 (high) Source edition 26-07 80.5
  6. 6 Gemini 3.7 Flash Source edition 26-07 80.5
  7. 7 Kimi K3 Source edition 26-07 79.1
  8. 8 Grok 4.6 Source edition 26-07 78.0
  9. 9 DeepSeek-V4-Pro Source edition 26-07 77.5
  10. 10 Kimi K2.6 Source edition 26-07 75.7
  11. 11 Claude Opus 4.8 Source edition 26-07 75.1
  12. 12 GLM-5.3 Source edition 26-07 74.4
  13. 13 DeepSeek-V4-Flash Source edition 26-07 73.7
  14. 14 Tencent Hy3 preview Source edition 26-07 73.6
  15. 15 MiniMax-M3 Source edition 26-07 69.0

Source: Official CompassRank leaderboard

#EntryScoreSettingDateSource
1 Claude Opus 5 83.3 Source edition 26-07 Jul 2026 Official leaderboard
2 GPT-5.6 Sol 80.8 Source edition 26-07 Jul 2026 Official leaderboard
3 GPT-5.5 80.6 Source edition 26-07 Apr 2026 Official leaderboard
4 Qwen 3.8 Max 80.6 Source edition 26-07 Aug 2026 Official leaderboard
5 Doubao-Seed-2.1-Pro-260628 (high) 80.5 Source edition 26-07 Official leaderboard
6 Gemini 3.7 Flash 80.5 Source edition 26-07 Aug 2026 Official leaderboard
7 Kimi K3 79.1 Source edition 26-07 Jul 2026 Official leaderboard
8 Grok 4.6 78.0 Source edition 26-07 Aug 2026 Official leaderboard
9 DeepSeek-V4-Pro 77.5 Source edition 26-07 Apr 2026 Official leaderboard
10 Kimi K2.6 75.7 Source edition 26-07 Apr 2026 Official leaderboard
11 Claude Opus 4.8 75.1 Source edition 26-07 May 2026 Official leaderboard
12 GLM-5.3 74.4 Source edition 26-07 Aug 2026 Official leaderboard
13 DeepSeek-V4-Flash 73.7 Source edition 26-07 Apr 2026 Official leaderboard
14 Tencent Hy3 preview 73.6 Source edition 26-07 Apr 2026 Official leaderboard
15 MiniMax-M3 69.0 Source edition 26-07 Jun 2026 Official leaderboard

4 models have held number one on the official leaderboard since its release in September 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

Moonshot AI OpenAI Anthropic
Own leaderboard Not tested since Aug 2026
Claude Opus 5 83.3

Source: Official CompassRank leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard
65.0 70.0 75.0 80.0 85.0 2026 65.0 70.0 75.0 80.0 85.0 2026

Source: Official CompassRank leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Other

All 43 in Other
  • Personal Superintelligence Index (Meta) In use Meta's composite of seven personal and knowledge-work evals: Agents' Last Exam, AutomationBench-AA, FocusBench, JobBench, GDP.pdf, VBT and DRACO.
  • Artificial Analysis Intelligence Index v5 In use Version 5 of Artificial Analysis's index: nine evaluations in four equal categories, most with private test sets, run three times each in its own agent harness. Gemini 4 Carbon 59.0
  • SAGE v1.1 (Vals) In use Models grade handwritten undergraduate math and computer science work against rubrics. Scored by class-balanced accuracy of each rubric check against expert grading. Claude Opus 5.5 91.3%
  • MTEB Fading Massive Text Embedding Benchmark: retrieval, clustering, classification across languages and modalities. sonar 42%
  • Anthropic ECI (AECI) Only in system cards Index that combines a basket of capability benchmarks into one capability score to track trends across Anthropic models. Claude Opus 5.5 169.4
  • SPEED-Bench Only in system cards Measures speculative decoding across varied prompts and domains. Reports the average acceptance length of draft tokens, so higher means faster generation at the same output. Nemotron 3 Ultra 5.15