Mathematical Reasoning on AIME24 (accuracy (%))

86.3Accuracy (AIME24)

Qwen3-Base (DICE-PC)

Updated 1mo ago

Evaluation Results

Method	Links
Qwen3-Base (DICE-PC) 2026.06		86.3
OpenAI o3 2026.06		85.8
Qwen3-235B-A22B 2026.06		84.9
Qwen3-Base (DICE-FT) 2026.06		84.7
Qwen3-Inst (SC@64) 2026.06		83.1
Qwen3 (thinking) 2026.06		81.4
Qwen3-A3B 2026.06		80.4
Qwen3-Inst (Few-shot CoT) 2026.06		80
OpenAI o3-mini 2026.06		79.6
QwQ 2026.06		79.5
Qwen3-Inst (ReAct) 2026.06		78.4
Qwen3-Inst (SC@64) 2026.06		75.4
Qwen3-Inst (ToT) 2026.06		75.1
Qwen3-Base Debate+Judge (FT) 2026.06		71.1
Llama 3.1-Base (DICE-PC) 2026.06		70.3
DeepSeek-R1 Distill 2026.06		70
Qwen3-Base Role-play (FT) 2026.06		69.9
Qwen3-Inst Debate+Judge (prompt) 2026.06		68.9
Qwen3-Base Debate (FT) 2026.06		68.7
Qwen3-Inst Role-play (prompt) 2026.06		67.3
Qwen3-Base Debate+Judge (FT) 2026.06		66.8
Qwen3-Inst Debate (prompt) 2026.06		65.8
Qwen3-Base Role-play (FT) 2026.06		65.1
Qwen3-Base Debate (FT) 2026.06		63.9
Qwen3-Inst (Zero-shot CoT) 2026.06		60.7
Qwen3-Base-DAPO 2026.06		47.7
Qwen3-Base-GRPO 2026.06		45.2
Qwen3-Base-DAPO 2026.06		41.3
Qwen3-Base-GRPO 2026.06		39.1
Qwen3-Base-PPO 2026.06		36.8
Qwen3-Base 2026.06		31.7
Qwen3-Base 2026.06		31
Qwen2.5-Base 2026.06		18.9
GPT-4o-mini 2026.06		8.1