Interactive Reasoning on Timely-Eval
34.9Zork1 ScoreGemini2.5-pro
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Gemini2.5-provariant=pro2026.01 | 34.9 | 50.7 | 34.1 | 71.9 | |
| GPT-5.1(medium)variant=medium2026.01 | 34.1 | 57.6 | 24.4 | 105 | |
| TimelyLM-8Bsize=8B2026.01 | 27.5 | 48.5 | 29.5 | 88.1 | |
| DeepSeek-V3.22026.01 | 24.9 | 48.7 | 15.9 | 63.2 | |
| Qwen3-32Bsize=32B2026.01 | 14.4 | 38.2 | 11.7 | 70 | |
| Qwen3-14Bsize=14B2026.01 | 9.8 | 34.9 | 9.5 | 50.5 | |
| Qwen3-8Bsize=8B2026.01 | 2.3 | 36 | 5.2 | 54.1 |