Benchmark Almanac

Xbench

Claude Opus 5.5 75.7 leads since September 2026
Not in model cards

Measures firsthand X posts about frontier models and coding harnesses. An LLM labels each post's sentiment, aggregated per model per week.

Every label is public by post id so numbers can be audited. The raw corpus stays private. The score is net firsthand sentiment in percentage points, from negative 100 to positive 100.

Released
September 2026
Kind
Other
Scored
Ranked by votes
Answers
Open-ended
Tasks
All public
Strongest model tested
GPT-6 Astra #3 on the capability index
Last model added
September 2026
Scores refreshed
November 12th, 2026
17
Models scored on its own leaderboard

Claude Opus 5.5 leads with 75.7 among models released by September 2026

Best score per model from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best.

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

17 models. Scroll sideways for the rest.

  1. 75.7 Claude Opus 5.5
  2. 58.1 Kimi K3
  3. 52.9 GLM-5.3-Flash
  4. 41.5 Gemini 3.7 Flash
  5. 40.7 Grok 4.6
  6. 40.2 GPT-6 Astra
  7. 39.5 GLM-5.3
  8. 36.7 GPT-5.6 Luna
  9. 35.2 Muse Spark 1.3
  10. 28.9 GPT-6 Luna
  11. 26.9 Muse Spark 1.2
  12. 24.9 Claude Fable 5.1
  13. 22.9 Gemini 3.8 Flash
  14. 18.2 GPT-5.6 Sol
  15. -11.11 Grok 4.7
  16. -16.96 GPT-6 Sol
  17. -31.93 Claude Opus 5
  1. 1 Claude Opus 5.5 75.7
  2. 2 Kimi K3 58.1
  3. 3 GLM-5.3-Flash 52.9
  4. 4 Gemini 3.7 Flash 41.5
  5. 5 Grok 4.6 40.7
  6. 6 GPT-6 Astra 40.2
  7. 7 GLM-5.3 39.5
  8. 8 GPT-5.6 Luna 36.7
  9. 9 Muse Spark 1.3 35.2
  10. 10 GPT-6 Luna 28.9
  11. 11 Muse Spark 1.2 26.9
  12. 12 Claude Fable 5.1 24.9
  13. 13 Gemini 3.8 Flash 22.9
  14. 14 GPT-5.6 Sol 18.2
  15. 15 Grok 4.7 -11.11
  16. 16 GPT-6 Sol -16.96
  17. 17 Claude Opus 5 -31.93

Source: Official Xbench leaderboard

#ModelScoreSettingDateSource
1 Claude Opus 5.5 75.7 Sep 2026 Official leaderboard
2 Kimi K3 58.1 Jul 2026 Official leaderboard
3 GLM-5.3-Flash 52.9 Aug 2026 Official leaderboard
4 Gemini 3.7 Flash 41.5 Aug 2026 Official leaderboard
5 Grok 4.6 40.7 Aug 2026 Official leaderboard
6 GPT-6 Astra 40.2 Sep 2026 Official leaderboard
7 GLM-5.3 39.5 Aug 2026 Official leaderboard
8 GPT-5.6 Luna 36.7 Jul 2026 Official leaderboard
9 Muse Spark 1.3 35.2 Sep 2026 Official leaderboard
10 GPT-6 Luna 28.9 Sep 2026 Official leaderboard
11 Muse Spark 1.2 26.9 Aug 2026 Official leaderboard
12 Claude Fable 5.1 24.9 Sep 2026 Official leaderboard
13 Gemini 3.8 Flash 22.9 Sep 2026 Official leaderboard
14 GPT-5.6 Sol 18.2 Jul 2026 Official leaderboard
15 Grok 4.7 -11.11 Sep 2026 Official leaderboard
16 GPT-6 Sol -16.96 Sep 2026 Official leaderboard
17 Claude Opus 5 -31.93 Jul 2026 Official leaderboard

2 models have held number one on the official leaderboard since its release in September 2026

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

Moonshot AI Anthropic
Own leaderboard Not tested since Sep 2026
Claude Opus 5.5 75.7

Source: Official Xbench 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
-50.00 0.00 50.0 100.0 -50.00 0.00 50.0 100.0

Source: Official Xbench 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