Iterative Self-Optimization on OPT-BENCH 5 steps 1.0 (ML environments)
8Win Countclaude-3-7-sonnet-20250219
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| claude-3-7-sonnet-20250219Steps=5, Model Category=Proprietary, Model Type=General2026.05 | 8 | 10 | 24 | 62 | 52 | |
| claude-3-5-sonnet-20241022Steps=5, Model Category=Proprietary, Model Type=General2026.05 | 7 | 20 | 32 | 49 | 44 | |
| Deepseek-V3.1-ThinkingSteps=5, Model Category=Proprietary, Model Type=Reasoning2026.05 | 6 | 0 | 0 | 76 | 74 | |
| Qwen3-235B-ThinkingSteps=5, Model Category=Proprietary, Model Type=Reasoning2026.05 | 6 | 0 | 0 | 74 | 76 | |
| Qwen3-235B-InstructSteps=5, Model Category=Proprietary, Model Type=General2026.05 | 6 | 16 | 15 | 48 | 50 | |
| Qwen2.5-3B-InstructSteps=5, Model Category=Open-Source, Model Type=General2026.05 | 6 | 64 | 80 | 12 | 3 | |
| Qwen2.5-7B-InstructSteps=5, Model Category=Open-Source, Model Type=General2026.05 | 6 | 54 | 70 | 22 | 14 | |
| gpt-4o-2024-08-06Steps=5, Model Category=Proprietary, Model Type=General2026.05 | 5 | 30 | 34 | 41 | 40 | |
| gpt-o3-miniSteps=5, Model Category=Proprietary, Model Type=Reasoning2026.05 | 5 | 0 | 0 | 72 | 68 | |
| gemini-2.0-flashSteps=5, Model Category=Proprietary, Model Type=General2026.05 | 5 | 16 | 20 | 40 | 42 | |
| grok-3Steps=5, Model Category=Proprietary, Model Type=General2026.05 | 5 | 16 | 14 | 52 | 55 | |
| Qwen2.5-14B-InstructSteps=5, Model Category=Open-Source, Model Type=General2026.05 | 5 | 40 | 60 | 31 | 22 | |
| Qwen2.5-32B-InstructSteps=5, Model Category=Open-Source, Model Type=General2026.05 | 5 | 42 | 44 | 33 | 35 | |
| Qwen2.5-72B-InstructSteps=5, Model Category=Open-Source, Model Type=General2026.05 | 5 | 40 | 50 | 33 | 30 | |
| Qwen3-8BSteps=5, Model Category=Open-Source, Model Type=General2026.05 | 5 | 28 | 22 | 51 | 56 | |
| Qwen3-30B-A3BSteps=5, Model Category=Open-Source, Model Type=General2026.05 | 5 | 4 | 12 | 60 | 59 | |
| Qwen3-32BSteps=5, Model Category=Open-Source, Model Type=General2026.05 | 5 | 14 | 12 | 52 | 57 | |
| gpt-4.1-2025-04-14Steps=5, Model Category=Proprietary, Model Type=General2026.05 | 4 | 10 | 10 | 45 | 47 | |
| Internlm3-8b-InstructSteps=5, Model Category=Open-Source, Model Type=General2026.05 | 4 | 52 | 54 | 12 | 14 |