Benchmark Almanac

MecEng

Claude Opus 4.8 91.4% leads since August 2026
Not in model cards

Models generate rigid and flexible multibody simulations from text. The evaluator checks system connections, geometry, and numerical outputs against engineering references.

The authors publish a rigid-body score workbook for 34 models. The full benchmark has 84 tasks, and the public runner skips withheld references.

Released
August 2026
Kind
Scientific code
Level
University
Scored
Checked automatically
Answers
Code
Tasks
Public, plus a private test set
Measures
Capability
Language
English
In and out
Text → Code
Interaction
Single turn
Environment
Code sandbox
Refresh
One fixed set
Run by
A host
Strongest model tested
Claude Opus 4.8 #22 on the capability index
Last model added
May 2026
Scores refreshed
November 20th, 2026
34
Entries on its own leaderboard, 20 models we track

Claude Opus 4.8 leads with 91.4% among models released by May 2026 Claude Opus 4.8 leads with 91.4% among models released by May 2026

Best score per entry from the benchmark's own leaderboard, read on November 20th, 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. This view shows only models we track, each at its best.

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

  1. 1 Claude Opus 4.8 Claude-Opus-4.8 / thinking False 91.4%
  2. 2 GPT-5.2 GPT-5.2 / thinking False 86%
  3. 3 Qwen3.6 27B Qwen3.6-27B-OL-T0.0 / thinking False 82.1%
  4. 4 Qwen3.5-27B Qwen3.5-27B-OL-T0.0 / thinking False 81.1%
  5. 5 Gemma 4 31B IT Gemma4-31B-OL-B2-T0.0 / thinking False 81.1%
  6. 6 gpt-oss-120b gpt-oss-120B-OL-B8-T0.0 / thinking low 73.3%
  7. 7 Qwen3.5-35B-A3B Qwen3.5-35B-OL-T0.0 / thinking False 70.3%
  8. 8 gpt-oss-20b gpt-oss-20B-OL-B32-T0.0 / thinking low 67.6%
  9. 9 Gemma 4 26B A4B Gemma4-26B-OL-B8-T0.0 / thinking False 66%
  10. 10 Qwen3.5-122B-A10B Qwen3.5-122B-OL-B8-T0.0 / thinking False 62.5%
  11. 11 Qwen 3.6 35B-A3B Qwen3.6-35B-OL-T0.0 / thinking False 60.7%
  12. 12 Qwen2.5-Coder-32B QwenCoder2.5-32B-OL-B4-T0.0 / thinking False 57.3%
  13. 13 Ministral 3 14B ministral-3-14B-OL-B16-T0.0 / thinking False 54.4%
  14. 14 Llama 3.3 70B Llama3.3-70B-OL-B16-T0.0 / thinking False 47.7%
  15. 15 Llama 3.1-70B Llama3.1-70B-OL-B16-T0.0 / thinking False 47.3%
  16. 16 Qwen2-72B Qwen2-72B-OL-B16-T0.0 / thinking False 38.7%
  17. 17 qwen3-coder:30b QwenCoder3-30B-OL-B8-T0.0 / thinking False 36.8%
  18. 18 phi4:latest Phi4-OL-B12-T0.0 / thinking False 36.7%
  19. 19 Qwen3.5-4B Qwen3.5-4B-OL-B32-T0.0 / thinking False 33.8%
  20. 20 Qwen3.5-9B Qwen3.5-9B-OL-B32-T0.0 / thinking False 33.2%
  21. 21 deepseek-r1:32b deepseek-r1-32B-OL-B2-T0.0 / thinking False 31.4%
  22. 22 glm-4.7-flash:q4_K_M glm-4.7-flash-30B-OL-B4-T0.0 / thinking False 29%
  23. 23 deepseek-r1:70b deepseek-r1-70B-OL-B16-T0.0 / thinking False 28.6%
  24. 24 nemotron-cascade-2:30b nemotron-cascade-2-30B-OL-B8-T0.0 / thinking False 16%
  25. 25 deepseek-coder-v2:latest DeepSeekCoder-16B-OL-B16-T0.0 / thinking False 12.3%
  26. 26 codegemma:7b codegemma-7B-OL-B12-T0.0 / thinking False 10.2%
  27. 27 dolphincoder:15b dolphincoder-15B-OL-B12-T0.0 / thinking False 10%
  28. 28 nemotron3:33b nemotron3-33B-OL-B3-T0.0 / thinking False 9.00%
  29. 29 Llama 3.1-8B Llama3.1-8B-OL-B24-T0.0 / thinking False 3.70%
  30. 30 codeqwen:7b CodeQwen-7B-OL-B24-T0.0 / thinking False 2.70%
  31. 31 Qwen2-7B Qwen2-7B-OL-B24-T0.0 / thinking False 2.00%
  32. 32 phi3:14b Phi3-OL-B16-T0.0 / thinking False 1.60%
  33. 33 qwen:72b Qwen-72B-OL-B4-T0.0 / thinking False 1.40%
  34. 34 qwen:32b Qwen-32B-OL-B8-T0.0 / thinking False 0%

Source: Official MecEng leaderboard

Source: Official MecEng leaderboard

#EntryScoreSettingReleasedRun onSource
1 Claude Opus 4.8 91.4% Claude-Opus-4.8 / thinking False, API May 2026 Jun 2026 Official leaderboard
2 GPT-5.2 86% GPT-5.2 / thinking False, API Dec 2025 Jun 2026 Official leaderboard
3 Qwen3.6 27B 82.1% Qwen3.6-27B-OL-T0.0 / thinking False, ollama Apr 2026 Jun 2026 Official leaderboard
4 Qwen3.5-27B 81.1% Qwen3.5-27B-OL-T0.0 / thinking False, ollama Feb 2026 Jun 2026 Official leaderboard
5 Gemma 4 31B IT 81.1% Gemma4-31B-OL-B2-T0.0 / thinking False, ollama Apr 2026 Jun 2026 Official leaderboard
6 gpt-oss-120b 73.3% gpt-oss-120B-OL-B8-T0.0 / thinking low, ollama Aug 2025 Jun 2026 Official leaderboard
7 Qwen3.5-35B-A3B 70.3% Qwen3.5-35B-OL-T0.0 / thinking False, ollama Feb 2026 Jun 2026 Official leaderboard
8 gpt-oss-20b 67.6% gpt-oss-20B-OL-B32-T0.0 / thinking low, ollama Aug 2025 Jun 2026 Official leaderboard
9 Gemma 4 26B A4B 66% Gemma4-26B-OL-B8-T0.0 / thinking False, ollama Apr 2026 Jun 2026 Official leaderboard
10 Qwen3.5-122B-A10B 62.5% Qwen3.5-122B-OL-B8-T0.0 / thinking False, ollama Feb 2026 Jun 2026 Official leaderboard
11 Qwen 3.6 35B-A3B 60.7% Qwen3.6-35B-OL-T0.0 / thinking False, ollama Apr 2026 Jun 2026 Official leaderboard
12 Qwen2.5-Coder-32B 57.3% QwenCoder2.5-32B-OL-B4-T0.0 / thinking False, ollama Sep 2024 Jun 2026 Official leaderboard
13 Ministral 3 14B 54.4% ministral-3-14B-OL-B16-T0.0 / thinking False, ollama Dec 2025 Jun 2026 Official leaderboard
14 Llama 3.3 70B 47.7% Llama3.3-70B-OL-B16-T0.0 / thinking False, ollama Dec 2024 Jun 2026 Official leaderboard
15 Llama 3.1-70B 47.3% Llama3.1-70B-OL-B16-T0.0 / thinking False, ollama Jul 2024 Jun 2026 Official leaderboard
16 Qwen2-72B 38.7% Qwen2-72B-OL-B16-T0.0 / thinking False, ollama Jun 2024 Jun 2026 Official leaderboard
17 qwen3-coder:30b 36.8% QwenCoder3-30B-OL-B8-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
18 phi4:latest 36.7% Phi4-OL-B12-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
19 Qwen3.5-4B 33.8% Qwen3.5-4B-OL-B32-T0.0 / thinking False, ollama Feb 2026 Jun 2026 Official leaderboard
20 Qwen3.5-9B 33.2% Qwen3.5-9B-OL-B32-T0.0 / thinking False, ollama Feb 2026 Jun 2026 Official leaderboard
21 deepseek-r1:32b 31.4% deepseek-r1-32B-OL-B2-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
22 glm-4.7-flash:q4_K_M 29% glm-4.7-flash-30B-OL-B4-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
23 deepseek-r1:70b 28.6% deepseek-r1-70B-OL-B16-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
24 nemotron-cascade-2:30b 16% nemotron-cascade-2-30B-OL-B8-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
25 deepseek-coder-v2:latest 12.3% DeepSeekCoder-16B-OL-B16-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
26 codegemma:7b 10.2% codegemma-7B-OL-B12-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
27 dolphincoder:15b 10% dolphincoder-15B-OL-B12-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
28 nemotron3:33b 9.00% nemotron3-33B-OL-B3-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
29 Llama 3.1-8B 3.70% Llama3.1-8B-OL-B24-T0.0 / thinking False, ollama Jul 2024 Jun 2026 Official leaderboard
30 codeqwen:7b 2.70% CodeQwen-7B-OL-B24-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
31 Qwen2-7B 2.00% Qwen2-7B-OL-B24-T0.0 / thinking False, ollama Jun 2024 Jun 2026 Official leaderboard
32 phi3:14b 1.60% Phi3-OL-B16-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
33 qwen:72b 1.40% Qwen-72B-OL-B4-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
34 qwen:32b 0% Qwen-32B-OL-B8-T0.0 / thinking False, ollama Jun 2026 Official leaderboard
#ModelScoreSettingReleasedRun onSource
1 Claude Opus 4.8 91.4% Claude-Opus-4.8 / thinking False, API May 2026 Jun 2026 Official leaderboard
2 GPT-5.2 86% GPT-5.2 / thinking False, API Dec 2025 Jun 2026 Official leaderboard
3 Qwen3.6 27B 82.1% Qwen3.6-27B-OL-T0.0 / thinking False, ollama Apr 2026 Jun 2026 Official leaderboard
4 Qwen3.5-27B 81.1% Qwen3.5-27B-OL-T0.0 / thinking False, ollama Feb 2026 Jun 2026 Official leaderboard
5 Gemma 4 31B IT 81.1% Gemma4-31B-OL-B2-T0.0 / thinking False, ollama Apr 2026 Jun 2026 Official leaderboard
6 gpt-oss-120b 73.3% gpt-oss-120B-OL-B8-T0.0 / thinking low, ollama Aug 2025 Jun 2026 Official leaderboard
7 Qwen3.5-35B-A3B 70.3% Qwen3.5-35B-OL-T0.0 / thinking False, ollama Feb 2026 Jun 2026 Official leaderboard
8 gpt-oss-20b 67.6% gpt-oss-20B-OL-B32-T0.0 / thinking low, ollama Aug 2025 Jun 2026 Official leaderboard
9 Gemma 4 26B A4B 66% Gemma4-26B-OL-B8-T0.0 / thinking False, ollama Apr 2026 Jun 2026 Official leaderboard
10 Qwen3.5-122B-A10B 62.5% Qwen3.5-122B-OL-B8-T0.0 / thinking False, ollama Feb 2026 Jun 2026 Official leaderboard
11 Qwen 3.6 35B-A3B 60.7% Qwen3.6-35B-OL-T0.0 / thinking False, ollama Apr 2026 Jun 2026 Official leaderboard
12 Qwen2.5-Coder-32B 57.3% QwenCoder2.5-32B-OL-B4-T0.0 / thinking False, ollama Sep 2024 Jun 2026 Official leaderboard
13 Ministral 3 14B 54.4% ministral-3-14B-OL-B16-T0.0 / thinking False, ollama Dec 2025 Jun 2026 Official leaderboard
14 Llama 3.3 70B 47.7% Llama3.3-70B-OL-B16-T0.0 / thinking False, ollama Dec 2024 Jun 2026 Official leaderboard
15 Llama 3.1-70B 47.3% Llama3.1-70B-OL-B16-T0.0 / thinking False, ollama Jul 2024 Jun 2026 Official leaderboard
16 Qwen2-72B 38.7% Qwen2-72B-OL-B16-T0.0 / thinking False, ollama Jun 2024 Jun 2026 Official leaderboard
17 Qwen3.5-4B 33.8% Qwen3.5-4B-OL-B32-T0.0 / thinking False, ollama Feb 2026 Jun 2026 Official leaderboard
18 Qwen3.5-9B 33.2% Qwen3.5-9B-OL-B32-T0.0 / thinking False, ollama Feb 2026 Jun 2026 Official leaderboard
19 Llama 3.1-8B 3.70% Llama3.1-8B-OL-B24-T0.0 / thinking False, ollama Jul 2024 Jun 2026 Official leaderboard
20 Qwen2-7B 2.00% Qwen2-7B-OL-B24-T0.0 / thinking False, ollama Jun 2024 Jun 2026 Official leaderboard

Number one over time

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

Anthropic
Own leaderboard Passed 85% in Aug 2026, by Claude Opus 4.8

Source: Official MecEng leaderboard

GPT-5.2 on the official leaderboard passed 85% in December 2025

Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Claude Opus 4.8 GPT-5.2 gpt-oss-120b Qwen2.5-Coder-32B Llama 3.1-70B Qwen2-72B 0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Claude Opus 4.8 gpt-oss-120b

Source: Official MecEng leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

Versions

More in Coding

All 523 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

Data audit

All 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access Public tasks, plus a private test set.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the official leaderboard (34 entries).
  • Pass Models scored 20 models we track, plus 14 other entries on its leaderboard.
  • Note Newest score The newest came 6 months ago, in May 2026, for Claude Opus 4.8. That’s expected for a saturated test.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Claude Opus 4.8 (#22).
  • Note Headroom Saturated. The best is 91.4%, past 85% since December 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 20 rows for models we track and 14 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check