Instruction Following on Helpsteer2 Trivial
78.22AccuracyDeepSeek-V3
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DeepSeek-V32026.01 | 78.22 | 87.8 | |
| Qwen3-30B-A3B-Thinking-2507Parameters=30B, Architecture=A3B, Type=Thinking2026.01 | 78.14 | 97.44 | |
| Qwen3-Next-80B-A4B-ThinkingParameters=80B, Architecture=A4B, Type=Thinking2026.01 | 77.94 | 91.18 | |
| QwQ-32BParameters=32B2026.01 | 76.49 | 91.11 | |
| Qwen3-Next-80B-A3B-InstructParameters=80B, Architecture=A3B, Type=Instruct2026.01 | 75.88 | 82.5 | |
| DeepSeek-R12026.01 | 73.61 | 95.24 | |
| Qwen3-30B-A3B-Instruct-2507Parameters=30B, Architecture=A3B, Type=Instruct2026.01 | 72.78 | 95.67 | |
| Qwen2.5-32B-InstructParameters=32B, Type=Instruct2026.01 | 71.13 | 83.19 |