Mathematical Reasoning on Avg. (held out)
62.8AccuracySC @ 64
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SC @ 64Base Model=Qwen3-8B, Type=Handcrafted2026.05 | 62.8 | 1,195.7 | |
| ASCBase Model=Qwen3-8B, Type=Handcrafted2026.05 | 62.8 | 485.7 | |
| ESCBase Model=Qwen3-8B, Type=Handcrafted2026.05 | 62.8 | 927.6 | |
| AutoTTS (β=1.0)Base Model=Qwen3-8B, Type=Discovered2026.05 | 62.7 | 710.8 | |
| Parallel-ProbeBase Model=Qwen3-8B, Type=Handcrafted2026.05 | 62 | 872 | |
| AutoTTS (β=0.5)Base Model=Qwen3-8B, Type=Discovered2026.05 | 61.1 | 379 | |
| AutoTTS (β=1.0)Base Model=Qwen3-4B, Type=Discovered2026.05 | 60.5 | 648.6 | |
| Parallel-ProbeBase Model=Qwen3-4B, Type=Handcrafted2026.05 | 60.4 | 839.2 | |
| ASCBase Model=Qwen3-4B, Type=Handcrafted2026.05 | 60.2 | 333.1 | |
| SC @ 64Base Model=Qwen3-4B, Type=Handcrafted2026.05 | 60.1 | 1,128.2 | |
| ESCBase Model=Qwen3-4B, Type=Handcrafted2026.05 | 60.1 | 891.8 | |
| AutoTTS (β=0.5)Base Model=Qwen3-4B, Type=Discovered2026.05 | 59.8 | 348.7 | |
| AutoTTS (β=1.0)Base Model=Qwen3-1.7B, Type=Discovered2026.05 | 40.6 | 646.1 | |
| AutoTTS (β=0.5)Base Model=Qwen3-1.7B, Type=Discovered2026.05 | 38.6 | 343.5 | |
| SC @ 64Base Model=Qwen3-1.7B, Type=Handcrafted2026.05 | 34.3 | 1,093.5 | |
| ASCBase Model=Qwen3-1.7B, Type=Handcrafted2026.05 | 34.3 | 593.6 | |
| ESCBase Model=Qwen3-1.7B, Type=Handcrafted2026.05 | 34.3 | 964.3 | |
| Parallel-ProbeBase Model=Qwen3-1.7B, Type=Handcrafted2026.05 | 33.7 | 818 | |
| AutoTTS (β=1.0)Base Model=Qwen3-0.6B, Type=Discovered2026.05 | 24.6 | 480.9 | |
| Parallel-ProbeBase Model=Qwen3-0.6B, Type=Handcrafted2026.05 | 24.1 | 716.2 | |
| SC @ 64Base Model=Qwen3-0.6B, Type=Handcrafted2026.05 | 23.2 | 914.2 | |
| ASCBase Model=Qwen3-0.6B, Type=Handcrafted2026.05 | 23.2 | 617.3 | |
| ESCBase Model=Qwen3-0.6B, Type=Handcrafted2026.05 | 23.2 | 896.4 | |
| AutoTTS (β=0.5)Base Model=Qwen3-0.6B, Type=Discovered2026.05 | 21.8 | 254 |