AI Engineering Lab model creation
Models generate Exudyn simulation code for 35 mechanical systems. The evaluator compares numerical outputs with reference simulations across ten parameter settings per system.
The authors publish results for 14 model configurations on 350 cases. Two configurations use different implementations of Llama 3.1 8B.
- Made by
- University of Innsbruck
- Released
- April 2025
- Kind
- Scientific code
- Level
- University
- Scored
- Checked automatically
- Answers
- Code
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Environment
- Code sandbox
- Items from
- Generated
- Refresh
- New versions replace it
- Run by
- Anyone
- Strongest model tested
-
Phi-4 #160 on the capability index
- Last model added
- January 2025
- Scores refreshed
- November 20th, 2026
Qwen2.5-Coder-32B-Instruct leads with 91.7% among models released by January 2025 Qwen2.5-Coder-32B-Instruct leads with 91.7% among models released by January 2025
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 January 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 91.7
Qwen2.5-Coder-32B-Instruct
- 85.4
Qwen2.5-72B
- 80.9
Llama 3.1-70B
- 78.9
Llama 3.3 70B Instruct
- 74.9
Phi-4
- 68.0
Codestral
- 64.6 Viper-Coder-v1.7-Vsm6.Q8_0.gguf
- 40.0 FluentlyLM-Prinum.Q4_K_M.gguf
- 34.9
Llama 3.1-8B
- 28.6
DeepSeek Coder 33B
- 26.3 DeepSeek-Coder-V2-Lite-Instruct
- 20.3
Qwen2.5-32B
- 19.7
Llama 3-8B
- 1
Qwen2.5-Coder-32B-Instruct q4_0 91.7%
- 2
Qwen2.5-72B HF 85.4%
- 3
Llama 3.1-70B HF 80.9%
- 4
Llama 3.3 70B Instruct HF 78.9%
- 5
Phi-4 Q4_0 74.9%
- 6
Codestral Q6_K 68%
- 7 Viper-Coder-v1.7-Vsm6.Q8_0.gguf Q8_0 64.6%
- 8 FluentlyLM-Prinum.Q4_K_M.gguf Q4_K_M 40%
- 9
Llama 3.1-8B HF 34.9%
- 10
DeepSeek Coder 33B Q4_0 28.6%
- 11 DeepSeek-Coder-V2-Lite-Instruct Q6_K 26.3%
- 12
Qwen2.5-32B Q4_K_L 20.3%
- 13
Llama 3-8B Q4_0 19.7%
Source: Official AI Engineering Lab model creation leaderboard
- 1
Qwen2.5-Coder-32B-Instruct q4_0 91.7%
- 2
Qwen2.5-72B HF 85.4%
- 3
Llama 3.1-70B HF 80.9%
- 4
Llama 3.3 70B Instruct HF 78.9%
- 5
Phi-4 Q4_0 74.9%
- 6
Codestral Q6_K 68%
- 7
Llama 3.1-8B HF 34.9%
- 8
DeepSeek Coder 33B Q4_0 28.6%
- 9 DeepSeek-Coder-V2-Lite-Instruct Q6_K 26.3%
- 10
Qwen2.5-32B Q4_K_L 20.3%
- 11
Llama 3-8B Q4_0 19.7%
Source: Official AI Engineering Lab model creation leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 91.7% | q4_0 | Nov 2024 | Official leaderboard |
| 2 | | 85.4% | HF | Sep 2024 | Official leaderboard |
| 3 | | 80.9% | HF | Jul 2024 | Official leaderboard |
| 4 | | 78.9% | HF | Dec 2024 | Official leaderboard |
| 5 | | 74.9% | Q4_0 | Dec 2024 | Official leaderboard |
| 6 | | 68% | Q6_K | Jan 2025 | Official leaderboard |
| 7 | Viper-Coder-v1.7-Vsm6.Q8_0.gguf | 64.6% | Q8_0 | Official leaderboard | |
| 8 | FluentlyLM-Prinum.Q4_K_M.gguf | 40% | Q4_K_M | Official leaderboard | |
| 9 | | 34.9% | HF | Jul 2024 | Official leaderboard |
| 10 | | 28.6% | Q4_0 | Nov 2023 | Official leaderboard |
| 11 | DeepSeek-Coder-V2-Lite-Instruct | 26.3% | Q6_K | Jun 2024 | Official leaderboard |
| 12 | | 20.3% | Q4_K_L | Sep 2024 | Official leaderboard |
| 13 | | 19.7% | Q4_0 | Apr 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 91.7% | q4_0 | Nov 2024 | Official leaderboard |
| 2 | | 85.4% | HF | Sep 2024 | Official leaderboard |
| 3 | | 80.9% | HF | Jul 2024 | Official leaderboard |
| 4 | | 78.9% | HF | Dec 2024 | Official leaderboard |
| 5 | | 74.9% | Q4_0 | Dec 2024 | Official leaderboard |
| 6 | | 68% | Q6_K | Jan 2025 | Official leaderboard |
| 7 | | 34.9% | HF | Jul 2024 | Official leaderboard |
| 8 | | 28.6% | Q4_0 | Nov 2023 | Official leaderboard |
| 9 | DeepSeek-Coder-V2-Lite-Instruct | 26.3% | Q6_K | Jun 2024 | Official leaderboard |
| 10 | | 20.3% | Q4_K_L | Sep 2024 | Official leaderboard |
| 11 | | 19.7% | Q4_0 | Apr 2024 | Official leaderboard |
Rankings by part
We read 35 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as flyingMassPoint, freeFallMassPoint and singleMassOscillator. Llama 3.1-70B leads 18 of them, and 6 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
flyingMassPoint
11 models we track have a score here.
- 1
DeepSeek Coder 33B ↑7 100%
- 2
Llama 3-8B ↑9 100%
- 3
Llama 3.1-70B 100%
- 4
Llama 3.1-8B ↑3 100%
- 5
Qwen2.5-72B ↓3 100%
- 6
Qwen2.5-Coder-32B-Instruct ↓5 100%
- 7
Llama 3.3 70B Instruct ↓3 100%
- 8
Phi-4 ↓3 100%
- 9
Codestral ↓3 100%
- 10
Qwen2.5-32B 90%
freeFallMassPoint
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-32B ↑8 100%
- 3
Qwen2.5-Coder-32B-Instruct ↓2 100%
- 4
Llama 3.3 70B Instruct 100%
- 5
Phi-4 100%
- 6
Codestral 100%
- 7
Llama 3.1-8B 90%
- 8
Qwen2.5-72B ↓6 70%
- 9
DeepSeek Coder 33B ↓1 50%
- 10
Llama 3-8B ↑1 20%
singleMassOscillator
11 models we track have a score here.
- 1
DeepSeek Coder 33B ↑7 100%
- 2 DeepSeek-Coder-V2-Lite-Instruct ↑7 100%
- 3
Llama 3.1-70B 100%
- 4
Qwen2.5-72B ↓2 100%
- 5
Qwen2.5-Coder-32B-Instruct ↓4 100%
- 6
Llama 3.3 70B Instruct ↓2 100%
- 7
Phi-4 ↓2 100%
- 8
Codestral ↓2 100%
- 9
Llama 3-8B ↑2 90%
- 10
Llama 3.1-8B ↓3 90%
singleMassOscillatorGravity
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Qwen2.5-Coder-32B-Instruct ↓2 100%
- 4
Llama 3.3 70B Instruct 100%
- 5
Phi-4 100%
- 6
Codestral 100%
- 7
Llama 3.1-8B 50%
- 8
Qwen2.5-32B ↑2 50%
- 9 DeepSeek-Coder-V2-Lite-Instruct 40%
- 10
DeepSeek Coder 33B ↓2 20%
sliderCrankSimple
11 models we track have a score here.
- 1
Qwen2.5-72B ↑1 100%
- 2
Llama 3.3 70B Instruct ↑2 70%
- 3
Codestral ↑3 50%
- 4
Qwen2.5-Coder-32B-Instruct ↓3 30%
- 5
Llama 3-8B ↑6 10%
- 6
Llama 3.1-70B ↓3 10%
- 7
Llama 3.1-8B 10%
- 8
Phi-4 ↓3 10%
- 9
DeepSeek Coder 33B ↓1 0%
- 10 DeepSeek-Coder-V2-Lite-Instruct ↓1 0%
singlePendulumElasticSpring
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Llama 3.3 70B Instruct ↑1 100%
- 4
Phi-4 ↑1 100%
- 5
Llama 3.1-8B ↑2 80%
- 6
Qwen2.5-Coder-32B-Instruct ↓5 60%
- 7
DeepSeek Coder 33B ↑1 20%
- 8
Qwen2.5-32B ↑2 10%
- 9
Llama 3-8B ↑2 0%
- 10 DeepSeek-Coder-V2-Lite-Instruct ↓1 0%
singleMassOscillatorUserFunction
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Qwen2.5-Coder-32B-Instruct ↓2 100%
- 4
Llama 3.3 70B Instruct 100%
- 5
DeepSeek Coder 33B ↑3 80%
- 6
Codestral 80%
- 7
Phi-4 ↓2 70%
- 8
Llama 3.1-8B ↓1 30%
- 9 DeepSeek-Coder-V2-Lite-Instruct 10%
- 10
Llama 3-8B ↑1 0%
spinningDisc
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Qwen2.5-Coder-32B-Instruct ↓2 100%
- 4
Llama 3.3 70B Instruct 100%
- 5
Codestral ↑1 100%
- 6 DeepSeek-Coder-V2-Lite-Instruct ↑3 90%
- 7
Phi-4 ↓2 60%
- 8
Qwen2.5-32B ↑2 20%
- 9
DeepSeek Coder 33B ↓1 0%
- 10
Llama 3-8B ↑1 0%
doubleMassOscillator
11 models we track have a score here.
- 1 DeepSeek-Coder-V2-Lite-Instruct ↑8 100%
- 2
Llama 3.1-70B ↑1 100%
- 3
Qwen2.5-72B ↓1 100%
- 4
Qwen2.5-Coder-32B-Instruct ↓3 100%
- 5
Llama 3.3 70B Instruct ↓1 100%
- 6
Phi-4 ↓1 100%
- 7
Codestral ↓1 100%
- 8
Llama 3-8B ↑3 90%
- 9
Llama 3.1-8B ↓2 90%
- 10
DeepSeek Coder 33B ↓2 60%
nMassOscillator
11 models we track have a score here.
- 1
Qwen2.5-72B ↑1 100%
- 2
Qwen2.5-Coder-32B-Instruct ↓1 100%
- 3
Llama 3.1-70B 90%
- 4
Llama 3.3 70B Instruct 90%
- 5
Phi-4 80%
- 6
Llama 3.1-8B ↑1 50%
- 7
DeepSeek Coder 33B ↑1 20%
- 8
Codestral ↓2 20%
- 9
Llama 3-8B ↑2 0%
- 10 DeepSeek-Coder-V2-Lite-Instruct ↓1 0%
singlePendulum
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Qwen2.5-Coder-32B-Instruct ↓2 90%
- 4
Phi-4 ↑1 80%
- 5
Llama 3.3 70B Instruct ↓1 60%
- 6
Codestral 20%
- 7
DeepSeek Coder 33B ↑1 0%
- 8
Llama 3-8B ↑3 0%
- 9 DeepSeek-Coder-V2-Lite-Instruct 0%
- 10
Llama 3.1-8B ↓3 0%
doublePendulum
11 models we track have a score here.
- 1
Phi-4 ↑4 100%
- 2
Qwen2.5-72B 90%
- 3
Qwen2.5-Coder-32B-Instruct ↓2 90%
- 4
Llama 3.1-8B ↑3 10%
- 5
DeepSeek Coder 33B ↑3 0%
- 6
Llama 3-8B ↑5 0%
- 7 DeepSeek-Coder-V2-Lite-Instruct ↑2 0%
- 8
Llama 3.1-70B ↓5 0%
- 9
Qwen2.5-32B ↑1 0%
- 10
Llama 3.3 70B Instruct ↓6 0%
nPendulum
11 models we track have a score here.
- 1
Llama 3.1-8B ↑6 100%
- 2
Qwen2.5-Coder-32B-Instruct ↓1 100%
- 3
Codestral ↑3 100%
- 4
Llama 3.3 70B Instruct 90%
- 5
Llama 3.1-70B ↓2 80%
- 6
Phi-4 ↓1 60%
- 7
DeepSeek Coder 33B ↑1 50%
- 8
Qwen2.5-72B ↓6 50%
- 9
Llama 3-8B ↑2 30%
- 10
Qwen2.5-32B 20%
fourBarMechanismPointMasses
11 models we track have a score here.
- 1
Qwen2.5-72B ↑1 100%
- 2
Qwen2.5-Coder-32B-Instruct ↓1 100%
- 3
Codestral ↑3 100%
- 4 DeepSeek-Coder-V2-Lite-Instruct ↑5 30%
- 5
Phi-4 20%
- 6
Qwen2.5-32B ↑4 10%
- 7
DeepSeek Coder 33B ↑1 0%
- 8
Llama 3-8B ↑3 0%
- 9
Llama 3.1-70B ↓6 0%
- 10
Llama 3.1-8B ↓3 0%
springCoupledFlyingRigidBodies
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Qwen2.5-Coder-32B-Instruct ↓2 100%
- 4
Llama 3.3 70B Instruct 100%
- 5
Phi-4 100%
- 6
Llama 3.1-8B ↑1 60%
- 7
DeepSeek Coder 33B ↑1 10%
- 8
Qwen2.5-32B ↑2 10%
- 9
Llama 3-8B ↑2 0%
- 10 DeepSeek-Coder-V2-Lite-Instruct ↓1 0%
torsionalOscillator
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Codestral ↑3 100%
- 4
Qwen2.5-Coder-32B-Instruct ↓3 80%
- 5
Phi-4 40%
- 6
Llama 3.1-8B ↑1 10%
- 7
DeepSeek Coder 33B ↑1 0%
- 8
Llama 3-8B ↑3 0%
- 9 DeepSeek-Coder-V2-Lite-Instruct 0%
- 10
Qwen2.5-32B 0%
invertedSinglePendulum
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Qwen2.5-Coder-32B-Instruct ↓2 100%
- 4
Llama 3.3 70B Instruct 100%
- 5
Codestral ↑1 80%
- 6
Llama 3.1-8B ↑1 40%
- 7
Phi-4 ↓2 20%
- 8 DeepSeek-Coder-V2-Lite-Instruct ↑1 10%
- 9
DeepSeek Coder 33B ↓1 0%
- 10
Llama 3-8B ↑1 0%
discRollingOnGround
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 80%
- 2
Llama 3.1-8B ↑5 80%
- 3
Qwen2.5-72B ↓1 80%
- 4
Qwen2.5-Coder-32B-Instruct ↓3 80%
- 5
Llama 3.3 70B Instruct ↓1 80%
- 6
Phi-4 ↓1 80%
- 7
Codestral ↓1 80%
- 8
Qwen2.5-32B ↑2 10%
- 9
DeepSeek Coder 33B ↓1 0%
- 10
Llama 3-8B ↑1 0%
doublePendulumElasticSpring
11 models we track have a score here.
- 1
Qwen2.5-Coder-32B-Instruct 80%
- 2
Qwen2.5-72B 70%
- 3
Phi-4 ↑2 40%
- 4 DeepSeek-Coder-V2-Lite-Instruct ↑5 20%
- 5
DeepSeek Coder 33B ↑3 10%
- 6
Llama 3.1-70B ↓3 10%
- 7
Llama 3.1-8B 10%
- 8
Llama 3.3 70B Instruct ↓4 10%
- 9
Llama 3-8B ↑2 0%
- 10
Qwen2.5-32B 0%
nPendulumElasticSpring
11 models we track have a score here.
- 1
Qwen2.5-72B ↑1 80%
- 2
Phi-4 ↑3 80%
- 3
Llama 3-8B ↑8 70%
- 4
Llama 3.1-8B ↑3 70%
- 5
Qwen2.5-Coder-32B-Instruct ↓4 70%
- 6
Llama 3.1-70B ↓3 60%
- 7
Llama 3.3 70B Instruct ↓3 60%
- 8
Codestral ↓2 50%
- 9
Qwen2.5-32B ↑1 20%
- 10
DeepSeek Coder 33B ↓2 10%
elasticChain
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Llama 3.3 70B Instruct ↑1 100%
- 4
Codestral ↑2 100%
- 5
Qwen2.5-Coder-32B-Instruct ↓4 80%
- 6
DeepSeek Coder 33B ↑2 30%
- 7
Qwen2.5-32B ↑3 30%
- 8
Llama 3.1-8B ↓1 20%
- 9
Llama 3-8B ↑2 10%
- 10 DeepSeek-Coder-V2-Lite-Instruct ↓1 10%
singlePendulumRigidBody
11 models we track have a score here.
- 1
DeepSeek Coder 33B ↑7 100%
- 2
Llama 3.1-70B ↑1 100%
- 3
Qwen2.5-72B ↓1 100%
- 4
Qwen2.5-Coder-32B-Instruct ↓3 100%
- 5
Llama 3.3 70B Instruct ↓1 100%
- 6
Phi-4 ↓1 100%
- 7
Codestral ↓1 100%
- 8
Llama 3.1-8B ↓1 80%
- 9 DeepSeek-Coder-V2-Lite-Instruct 50%
- 10
Llama 3-8B ↑1 0%
massPointOnStringRigid
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-Coder-32B-Instruct ↓1 100%
- 3
Llama 3.3 70B Instruct ↑1 100%
- 4
Phi-4 ↑1 100%
- 5
Qwen2.5-32B ↑5 80%
- 6
Codestral 60%
- 7
DeepSeek Coder 33B ↑1 0%
- 8
Llama 3-8B ↑3 0%
- 9 DeepSeek-Coder-V2-Lite-Instruct 0%
- 10
Llama 3.1-8B ↓3 0%
massPointOnStringElastic
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-Coder-32B-Instruct ↓1 100%
- 3
Llama 3.3 70B Instruct ↑1 100%
- 4
Phi-4 ↑1 100%
- 5
Codestral ↑1 40%
- 6
Qwen2.5-32B ↑4 30%
- 7
DeepSeek Coder 33B ↑1 0%
- 8
Llama 3-8B ↑3 0%
- 9 DeepSeek-Coder-V2-Lite-Instruct 0%
- 10
Llama 3.1-8B ↓3 0%
linkOnTwoPrismaticJoints
11 models we track have a score here.
- 1
Qwen2.5-72B ↑1 100%
- 2
Qwen2.5-Coder-32B-Instruct ↓1 100%
- 3
Llama 3.3 70B Instruct ↑1 100%
- 4
Llama 3.1-70B ↓1 90%
- 5
Codestral ↑1 90%
- 6
Phi-4 ↓1 60%
- 7
Qwen2.5-32B ↑3 30%
- 8
Llama 3-8B ↑3 10%
- 9
DeepSeek Coder 33B ↓1 0%
- 10 DeepSeek-Coder-V2-Lite-Instruct ↓1 0%
flyingRigidBody
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Qwen2.5-Coder-32B-Instruct ↓2 100%
- 4
Phi-4 ↑1 90%
- 5 DeepSeek-Coder-V2-Lite-Instruct ↑4 80%
- 6
Qwen2.5-32B ↑4 80%
- 7
Codestral ↓1 60%
- 8
Llama 3-8B ↑3 50%
- 9
DeepSeek Coder 33B ↓1 30%
- 10
Llama 3.3 70B Instruct ↓6 30%
suspendedRigidBody
11 models we track have a score here.
- 1
Qwen2.5-72B ↑1 100%
- 2
Qwen2.5-Coder-32B-Instruct ↓1 100%
- 3
Phi-4 ↑2 100%
- 4
Codestral ↑2 100%
- 5
Llama 3.3 70B Instruct ↓1 90%
- 6
Llama 3.1-70B ↓3 60%
- 7
Llama 3.1-8B 10%
- 8
DeepSeek Coder 33B 0%
- 9
Llama 3-8B ↑2 0%
- 10 DeepSeek-Coder-V2-Lite-Instruct ↓1 0%
gyroscopeOnSphericalJoint
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Llama 3.1-8B ↑5 100%
- 3
Llama 3.3 70B Instruct ↑1 100%
- 4
Phi-4 ↑1 100%
- 5
Codestral ↑1 100%
- 6
Llama 3-8B ↑5 90%
- 7 DeepSeek-Coder-V2-Lite-Instruct ↑2 80%
- 8
Qwen2.5-32B ↑2 60%
- 9
Qwen2.5-Coder-32B-Instruct ↓8 50%
- 10
Qwen2.5-72B ↓8 40%
prismaticJointSystem
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Llama 3.1-8B ↑5 100%
- 3
Qwen2.5-72B ↓1 100%
- 4
Qwen2.5-Coder-32B-Instruct ↓3 100%
- 5
Llama 3.3 70B Instruct ↓1 100%
- 6
Phi-4 ↓1 100%
- 7
Codestral ↓1 100%
- 8
DeepSeek Coder 33B 80%
- 9 DeepSeek-Coder-V2-Lite-Instruct 20%
- 10
Llama 3-8B ↑1 0%
twoMassPointsWithSprings
11 models we track have a score here.
- 1
DeepSeek Coder 33B ↑7 100%
- 2
Llama 3.1-70B ↑1 100%
- 3
Qwen2.5-72B ↓1 100%
- 4
Qwen2.5-Coder-32B-Instruct ↓3 100%
- 5
Llama 3.3 70B Instruct ↓1 100%
- 6
Phi-4 ↓1 100%
- 7
Codestral ↓1 100%
- 8
Llama 3-8B ↑3 20%
- 9
Llama 3.1-8B ↓2 10%
- 10 DeepSeek-Coder-V2-Lite-Instruct ↓1 0%
twoMassPointsWithDistances
11 models we track have a score here.
- 1 DeepSeek-Coder-V2-Lite-Instruct ↑8 100%
- 2
Llama 3.1-70B ↑1 100%
- 3
Qwen2.5-72B ↓1 100%
- 4
Qwen2.5-Coder-32B-Instruct ↓3 100%
- 5
Llama 3.3 70B Instruct ↓1 100%
- 6
Phi-4 ↓1 100%
- 7
Codestral ↓1 80%
- 8
DeepSeek Coder 33B 60%
- 9
Qwen2.5-32B ↑1 30%
- 10
Llama 3-8B ↑1 20%
rigidRotorSimplySupported
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Qwen2.5-72B 100%
- 3
Qwen2.5-Coder-32B-Instruct ↓2 100%
- 4
Llama 3.3 70B Instruct 100%
- 5
Llama 3.1-8B ↑2 50%
- 6
Phi-4 ↓1 40%
- 7
DeepSeek Coder 33B ↑1 0%
- 8
Llama 3-8B ↑3 0%
- 9 DeepSeek-Coder-V2-Lite-Instruct 0%
- 10
Qwen2.5-32B 0%
rigidRotorUnbalanced
11 models we track have a score here.
- 1
Qwen2.5-72B ↑1 100%
- 2
Qwen2.5-Coder-32B-Instruct ↓1 100%
- 3
Llama 3.3 70B Instruct ↑1 80%
- 4
Llama 3.1-70B ↓1 50%
- 5
DeepSeek Coder 33B ↑3 0%
- 6
Llama 3-8B ↑5 0%
- 7 DeepSeek-Coder-V2-Lite-Instruct ↑2 0%
- 8
Llama 3.1-8B ↓1 0%
- 9
Qwen2.5-32B ↑1 0%
- 10
Phi-4 ↓5 0%
doublePendulumRigidBodies
11 models we track have a score here.
- 1
Llama 3.1-70B ↑2 100%
- 2
Llama 3.1-8B ↑5 100%
- 3
Qwen2.5-72B ↓1 100%
- 4
Qwen2.5-Coder-32B-Instruct ↓3 100%
- 5
Llama 3.3 70B Instruct ↓1 100%
- 6
Phi-4 ↓1 100%
- 7
Codestral ↓1 100%
- 8 DeepSeek-Coder-V2-Lite-Instruct ↑1 90%
- 9
Llama 3-8B ↑2 60%
- 10
DeepSeek Coder 33B ↓2 50%
sliderCrankRigidBodies
11 models we track have a score here.
- 1
Qwen2.5-Coder-32B-Instruct 100%
- 2
Phi-4 ↑3 90%
- 3
Codestral ↑3 70%
- 4
Qwen2.5-72B ↓2 10%
- 5
DeepSeek Coder 33B ↑3 0%
- 6
Llama 3-8B ↑5 0%
- 7 DeepSeek-Coder-V2-Lite-Instruct ↑2 0%
- 8
Llama 3.1-70B ↓5 0%
- 9
Llama 3.1-8B ↓2 0%
- 10
Qwen2.5-32B 0%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official AI Engineering Lab model creation leaderboard
Qwen2.5-72B on the official leaderboard passed 85% in September 2024
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.
Source: Official AI Engineering Lab model creation 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
- 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 All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
- Pass Independent runs Scores from the official leaderboard (13 entries).
- Pass Models scored 11 models we track, plus 2 other entries on its leaderboard.
- Note Newest score The newest came 22 months ago, in January 2025, for Codestral. That’s expected for a test replaced by MecEng.
- Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Phi-4 (#160).
- Note Headroom Saturated. The best is 91.7%, past 85% since September 2024.
Leaderboard import 3 pass
- Pass Importer We read the board on github.com: 12 rows for models we track and 2 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.