List your company

Mathematical Reasoning on GSM8K (evaluation)

93.3AccuracyReasoning
83.10885.75488.491.046Jan 28, 2026

Evaluation results

MethodLinks
ReasoningModel=Qwen3-8B2026.0193.31,632—
KV-ClassificationModel=Qwen3-8B2026.0191.4554-66.1
Fast ThinkingModel=Qwen3-8B2026.0190.4211—
KV-GenerativeModel=Qwen3-8B2026.0189.2273-83.3
ReasoningModel=DeepSeek-R1-14B2026.0184.7432—
Fast ThinkingModel=DeepSeek-R1-14B2026.0184.5218—
KV-ClassificationModel=DeepSeek-R1-14B2026.0184.5218-49.5
KV-GenerativeModel=DeepSeek-R1-14B2026.0183.5242-44