List your company

Instruction Following on Helpsteer2 Trivial

78.22AccuracyDeepSeek-V3
70.846472.760774.67576.5893Jan 7, 2026

Evaluation results

MethodLinks
DeepSeek-V32026.0178.2287.8
Qwen3-30B-A3B-Thinking-2507Parameters=30B, Architecture=A3B, Type=Thinking2026.0178.1497.44
Qwen3-Next-80B-A4B-ThinkingParameters=80B, Architecture=A4B, Type=Thinking2026.0177.9491.18
QwQ-32BParameters=32B2026.0176.4991.11
Qwen3-Next-80B-A3B-InstructParameters=80B, Architecture=A3B, Type=Instruct2026.0175.8882.5
DeepSeek-R12026.0173.6195.24
Qwen3-30B-A3B-Instruct-2507Parameters=30B, Architecture=A3B, Type=Instruct2026.0172.7895.67
Qwen2.5-32B-InstructParameters=32B, Type=Instruct2026.0171.1383.19