General Language Modeling on CodaSet OOD Average (test)
87.84Performance (%)Qwen3-235B
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen3-235B2026.05 | 87.84 | 6.4 | |
| DeepSeek-V3.12026.05 | 87.59 | 7.1 | |
| DeepSeek-V3.2-Exp2026.05 | 87 | 6.8 | |
| DecoRBase Model=DeepSeek-V3.12026.05 | 86.49 | 5 | |
| gpt-oss-120b2026.05 | 84.96 | 3.8 | |
| Kimi-K2-Instruct2026.05 | 84.52 | 14.9 | |
| EmbedLLM2026.05 | 83.72 | 5.4 | |
| DecoRBase Model=gpt-oss-120b2026.05 | 83.67 | 3.3 | |
| MODEL-SAT2026.05 | 83.6 | 3.3 | |
| gemma-3-27b-it2026.05 | 83.25 | 1.3 | |
| RouterDC2026.05 | 82.77 | 4.2 | |
| Random Router2026.05 | 82.47 | 4.9 | |
| LLM Router2026.05 | 81.7 | 1.8 | |
| KNN Router2026.05 | 81.53 | 1.4 | |
| gemma-3-12b-it2026.05 | 81.42 | 1 | |
| Mistral-3.2-24B2026.05 | 81.11 | 2.2 |