Benchmark Almanac

OWE-Bench

Claude Fable 5.2 54.6% leads since November 2026
Not in model cards

26 long-horizon engineering and research projects across software, physical engineering, machine learning and scientific computing, from GPU kernels to aircraft design. Agents get 4 to 18 hours per attempt, and an offline verifier scores what they ship on held-out cases.

Each task is rescaled so the untouched baseline scores 0% and the task's ceiling scores 100%, and the board reports the mean of eight attempts. Mercor set exploits that raised a score to zero. Ten of the 26 graders stay private because they contain hidden data.

Made by
Mercor
Released
November 2026
Kind
Building whole projects
Level
Research frontier
Scored
Checked automatically
Answers
Agent task
Tasks
Public, plus a private test set
Strongest model tested
Claude Fable 5.2 #1 on the capability index
Last model added
November 2026
Scores refreshed
November 12th, 2026
14
Models scored on its own leaderboard

Claude Fable 5.2 leads with 54.6%

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.

14 models. Scroll sideways for the rest.

  1. 54.6 Claude Fable 5.2
  2. 44.2 GPT-6 Astra
  3. 41.6 Claude Opus 5.5
  4. 41.2 GPT-6.1 Sol
  5. 39.0 Claude Sonnet 5.5
  6. 32.8 Gemini 4 Carbon
  7. 29.1 Claude Fable 5.1
  8. 24.1 Gemini 4 Argon
  9. 13.6 Kimi K3.1 Pro
  10. 12.4 DeepSeek V4.1 Flash
  11. 11.8 GLM-5.3
  12. 7.00 Grok 4.7
  13. 3.30 Beam 1.0
  14. 1.10 Inkling
  1. 1 Claude Fable 5.2 max 54.6%
  2. 2 GPT-6 Astra max 44.2%
  3. 3 Claude Opus 5.5 max 41.6%
  4. 4 GPT-6.1 Sol max 41.2%
  5. 5 Claude Sonnet 5.5 max 39%
  6. 6 Gemini 4 Carbon high 32.8%
  7. 7 Claude Fable 5.1 max 29.1%
  8. 8 Gemini 4 Argon high 24.1%
  9. 9 Kimi K3.1 Pro max 13.6%
  10. 10 DeepSeek V4.1 Flash max 12.4%
  11. 11 GLM-5.3 max 11.8%
  12. 12 Grok 4.7 xhigh 7.00%
  13. 13 Beam 1.0 max 3.30%
  14. 14 Inkling max 1.10%

Source: Official OWE-Bench leaderboard

#ModelScoreSettingDateSource
1 Claude Fable 5.2 54.6% max Nov 2026 Official leaderboard
2 GPT-6 Astra 44.2% max Sep 2026 Official leaderboard
3 Claude Opus 5.5 41.6% max Sep 2026 Official leaderboard
4 GPT-6.1 Sol 41.2% max Sep 2026 Official leaderboard
5 Claude Sonnet 5.5 39% max Sep 2026 Official leaderboard
6 Gemini 4 Carbon 32.8% high Oct 2026 Official leaderboard
7 Claude Fable 5.1 29.1% max Sep 2026 Official leaderboard
8 Gemini 4 Argon 24.1% high Sep 2026 Official leaderboard
9 Kimi K3.1 Pro 13.6% max Oct 2026 Official leaderboard
10 DeepSeek V4.1 Flash 12.4% max Sep 2026 Official leaderboard
11 GLM-5.3 11.8% max Aug 2026 Official leaderboard
12 Grok 4.7 7.00% xhigh Sep 2026 Official leaderboard
13 Beam 1.0 3.30% max Oct 2026 Official leaderboard
14 Inkling 1.10% max Jul 2026 Official leaderboard

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

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

OpenAI Anthropic
Own leaderboard 2 leaders
Claude Fable 5.2 54.6%

Source: Official OWE-Bench 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 Best so far
0% 20% 40% 60% 80% 100% 2027 Near the ceiling (85%) Claude Fable 5.2 GPT-6 Astra Claude Fable 5.1 GLM-5.3 Inkling 0% 20% 40% 60% 80% 100% 2027 Near the ceiling (85%) Claude Fable 5.2 GPT-6 Astra

Source: Official OWE-Bench 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 Coding

All 282 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471