| Benchmark | Model | Rank | Results |
|---|---|---|---|
| code-generation-on-livecodebench-v6 | GlimpRouter (Qwen3-4B SLM + DeepSeek-R1-Distill-Qwen-32B LLM) | #56 | Pass@1: 43.57 |
| mathematical-reasoning-on-aime24 | GlimpRouter (Qwen3-4B SLM + DeepSeek-R1-Distill-Qwen-32B LLM) | #78 | Acc: 51.67 |
| mathematical-reasoning-on-aime-2025 | GlimpRouter (Qwen3-4B SLM + DeepSeek-R1-Distill-Qwen-32B LLM) | #99 | Accuracy: 39.17 |