Engineering Design Workflow Evaluation on Beams2D
100TCGPT-5-mini
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GPT-5-miniStyle=FULL2026.05 | 100 | 65 | |
| GPT-5-miniStyle=W-RAND2026.05 | 100 | 66 | |
| GPT-5-miniStyle=W-DERIVED2026.05 | 100 | 65 | |
| GPT-5-miniStyle=W-DISTRACT2026.05 | 100 | 63 | |
| Gemini-3-FlashStyle=NATURAL2026.05 | 100 | 88 | |
| Gemini-3-FlashStyle=W-RAND2026.05 | 100 | 65 | |
| Gemini-3-FlashStyle=W-DERIVED2026.05 | 100 | 64 | |
| Gemini-3-FlashStyle=W-DISTRACT2026.05 | 100 | 63 | |
| Gemini-3-FlashStyle=W-MULTI2026.05 | 100 | 70 | |
| Qwen3-4BStyle=W-RAND2026.05 | 100 | 64 | |
| Qwen3-4BStyle=W-DISTRACT2026.05 | 100 | 61 | |
| Qwen3-4BStyle=W-MULTI2026.05 | 100 | 65 | |
| Qwen3.5-4BStyle=W-RAND2026.05 | 100 | 69 | |
| Qwen3.5-4BStyle=W-MULTI2026.05 | 100 | 69 | |
| GPT-5-miniStyle=W-COND2026.05 | 93 | 66 | |
| GPT-5-miniStyle=W-MULTI2026.05 | 93 | 66 | |
| Gemini-3-FlashStyle=FULL2026.05 | 93 | 69 | |
| Qwen3.5-4BStyle=W-DISTRACT2026.05 | 93 | 66 | |
| GPT-5-miniStyle=NATURAL2026.05 | 87 | 82 | |
| Gemini-3-FlashStyle=W-COND2026.05 | 87 | 65 | |
| Qwen3.5-4BStyle=W-DERIVED2026.05 | 85 | 66 | |
| Qwen3.5-4BStyle=FULL2026.05 | 73 | 59 | |
| Qwen3.5-4BStyle=W-COND2026.05 | 60 | 62 | |
| Qwen3-4BStyle=W-DERIVED2026.05 | 47 | 53 | |
| Qwen3-4BStyle=W-COND2026.05 | 40 | 57 | |
| Qwen3.5-4BStyle=NATURAL2026.05 | 33 | 48 | |
| GPT-5-miniWorkflow Style=W-RAND2026.05 | 1 | 0.66 | |
| GPT-5-miniWorkflow Style=W-DISTRACT2026.05 | 1 | 0.63 | |
| Gemini-3-FlashWorkflow Style=W-RAND2026.05 | 1 | 0.65 | |
| Gemini-3-FlashWorkflow Style=W-DISTRACT2026.05 | 1 | 0.63 | |
| Qwen3-4BWorkflow Style=W-RAND2026.05 | 1 | 0.64 | |
| Qwen3-4BWorkflow Style=W-DISTRACT2026.05 | 1 | 0.61 | |
| Qwen3.5-4BWorkflow Style=W-RAND2026.05 | 1 | 0.69 | |
| GPT-5-miniWorkflow Style=W-COND2026.05 | 0.93 | 0.66 | |
| Qwen3.5-4BWorkflow Style=W-DISTRACT2026.05 | 0.93 | 0.66 | |
| Gemini-3-FlashWorkflow Style=W-COND2026.05 | 0.87 | 0.65 | |
| Qwen3.5-4BWorkflow Style=W-COND2026.05 | 0.6 | 0.62 | |
| Qwen3-4BWorkflow Style=W-COND2026.05 | 0.4 | 0.57 | |
| Qwen3-4BStyle=FULL2026.05 | 0 | 45 | |
| Qwen3-4BStyle=NATURAL2026.05 | 0 | 29 |