List your company

Mathematical Problem Solving on IneqMath (IM)

76Exact Match AccuracyDeepMath (vanilla)
13.629.84662.2Feb 1, 2026

Evaluation results

MethodLinks
DeepMath (vanilla)Optimization=vanilla2026.0276
DeepMath (GRPO)Optimization=GRPO2026.0273
DeepMath (RULES)Optimization=RULES2026.0271
Qwen2.5 (vanilla)Optimization=vanilla2026.0255
Qwen2.5 (GRPO)Optimization=GRPO2026.0254
Qwen2.5 (RULES)Optimization=RULES2026.0253
OLMO2 (vanilla)Optimization=vanilla2026.0229
OLMO2 (RULES)Optimization=RULES2026.0224
OLMO2 (GRPO)Optimization=GRPO2026.0223
Llama-3.1 (vanilla)Optimization=vanilla2026.0222
Llama-3.1 (GRPO)Optimization=GRPO2026.0218
Llama-3.1 (RULES)Optimization=RULES2026.0216