Share your thoughts, 1 month free Claude Pro on usSee more

Mathematical Reasoning on AIME 2025 (Pass@1, #Tokens)

93.3Pass@1 Accuracy

GPT-5(high)

Updated 1mo ago

Evaluation Results

Method
GPT-5(high) 2026.06	93.3	-
DeepSeek-V4-pro 2026.06	93.3	-
DeepSeek-R1 2026.06	90	-
Qwen3-max 2026.06	80	-
REBALANCE 2026.03	76.7	9,417
Baseline 2026.03	73.3	14,552
Qwen3-32B 2026.06	73.3	-
DiScOw/o ITDE 2026.06	66.7	-
DiScO 2026.06	66.7	-
DeepSeek-R1-Distill-Qwen-32B 2026.06	60	-
TOPD 2026.05	53.3	-
GPT-o1-mini 2026.06	50.8	-
DiScOw/o ITDE 2026.06	50	-
DiScO 2026.06	50	-
OPD 2026.05	46.7	-
GPT-o1-preview 2026.06	46.7	-
Qwen3-8B 2026.06	46.7	-
rStar-Math-7B 2026.06	46.7	-
QwQ-32B-preview 2026.06	46.7	-
ICPO 2025.10	43.7	-
Claude3.5-Sonnet 2026.06	43.3	-
ICPO† 2025.10	40.9	-
GRPOExpertDomain 2025.10	40.4	-
GRPOExtraRollouts 2025.10	40	-
Qwen3-4B 2026.05	40	-
DeepSeek-V3 2026.06	39.2	-
GRPO 2025.10	38.5	-
ReasonFlux-32B 2026.06	37.2	-
ReasonFlux-7B 2026.06	36.7	-
Sky-T1-32B-preview 2026.06	36.7	-
DeepSearch-1.5B 2025.09	35.42	-
Nemotron-Research-Reasoning-Qwen-1.5B v1 2025.09	33.85	-
Nemotron-Research-Reasoning-Qwen-1.5B v2 2025.09	32.92	-
DeepScaleR-1.5B 2025.09	30.52	-
LoRA-α 2026.06	28.95	-
Full Fine-Tuning 2026.06	28.69	-
LoRA (10× LR) 2026.06	26.71	-
GPT-4o 2026.06	26.7	-
DeepSeek-Coder-V2-Instruct 2026.06	26.7	-
Qwen2.5-Math-72B-Instruct 2026.06	26.7	-
ICPO† 2025.10	26.6	-
ICPO 2025.10	26.3	-
LoRA (10× LR) 2026.06	25.83	-
LoRA-α 2026.06	25.83	-
STILL-3-1.5B 2025.09	25	-
GRPOExpertDomain 2025.10	24.8	-
Open-RS3-1.5B 2025.09	24.79	-
GRPOExtraRollouts 2025.10	24.7	-
Open-RS2-1.5B 2025.09	24.37	-
DeepSeek-R1-Distill-Qwen-1.5B 2025.09	24.06	-
LoRA 2026.06	23.69	-
Open-RS1-1.5B 2025.09	22.6	-
GRPO 2025.10	22.5	-
S2L-PO 2026.05	22.5	-
DIVER 2026.06	22.5	-
Qwen3 2025.10	21.4	-
Qwen3 2025.10	20.6	-
LoRA 2026.06	19.73	-
Qwen2.5-Math-7B-Instruct 2026.06	16.7	-
GRPO w/ Clip-higher 2026.06	16.7	-
Pass@k Training 2026.06	16.6	-
S2L-PO 2026.05	14.6	-
NuminaMath-72B-CoT 2026.06	13.3	-
SuperCorrect-7B 2026.06	13.3	-
DeepSeek-R1-Distill-Qwen-7B 2026.06	13.3	-
Entropy-RL 2026.06	13.3	-
Qwen2.5-32B-Instruct 2026.06	13.3	-
GRPO 2026.05	12.9	-
GRPO 2026.05	12.1	-
Qwen2.5-Math-1.5B-Oat-Zero 2025.09	10	-
LLaMA3.1-405B-Instruct 2026.06	10	-
Qwen2.5-Math-1.5B-Instruct 2025.09	8.85	-
Qwen2.5-Math-1.5B 2025.09	6.35	-
S2L-PO 2026.05	3.5	-
LLaMA3.1-70B-Instruct 2026.06	3.3	-
Qwen2.5-Math-7B 2026.06	3.3	-
GRPO 2026.05	0.1	-
LLaMA3.1-8B-Instruct 2026.06	0	-
Mathstral-7B-v0.1 2026.06	0	-