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

Mathematical Reasoning on AIME 2025 (Accuracy and Average Tokens)

96.7Accuracy

PC-cubic

Updated 21h ago

Evaluation Results

Method
PC-cubic 2026.05	96.7	-
GPT-5-High 2026.05	94.6	-
SU-01 2026.05	94.6	-
Nemotron-Cascade-2 2026.05	94.2	-
PC-cubic 2026.05	94.1	-
Standard MV 2026.05	93.5	-
DeepSeek-V3.2 2026.05	93.1	-
Qwen3.6-35B-A3B 2026.05	92.5	-
Qwen3-235B-A22B-Thinking-2507 2026.05	92.3	-
ZAYA1-8B + Markovian RSA (40K/4K) 2026.05	91.9	-
GLM-4.7-Flash 2026.05	91.3	-
P1-30B-A3B 2026.05	90.4	-
Standard MV 2026.05	90.1	-
Gemma-4-31B 2026.05	88.8	-
ZAYA1-8B (single rollout) 2026.05	88.3	-
Gemini-2.5 Pro 2026.05	88	-
DeepSeek-R1-0528 2026.05	87.5	-
RecursiveMAS 2026.04	86.7	-
Full Attn 2026.05	86.67	-
RTPurbo 2026.05	86.67	-
BCR-Qwen3-4B (Ours) 2026.04	83.3	17,498
Upper Bound 2025.08	82.22	-
LaTER (training) 2026.05	80	10,575
RTPurbo 2026.05	80	-
Step-GRPO 2026.04	73.3	12,859
Single Agent 2026.04	73.3	-
TextGrad 2026.04	73.3	-
Recursive-TextMAS 2026.04	73.3	-
CoT-SFT 2026.05	73.3	12,687
LaTER (training-free) 2026.05	73.3	10,661
Qwen3-4B-Thinking-2507 2026.04	70	20,773
Single Agent 2026.04	70	-
Upper Bound 2025.08	70	-
CoT Baseline 2026.05	70	15,730
PiCSAR 2025.08	68.89	-
Average 2025.08	67.04	-
GRPO 2026.04	66.7	15,407
GRPO+SOP 2026.04	66.7	11,844
LoopLM 2026.04	66.7	-
Self-Consistency 2025.08	65.56	-
GRPO-λ 2026.04	63.3	12,321
Vanilla 2026.04	63.3	18,049
GRPO 2026.04	63.3	14,249
GRPO+LP 2026.04	63.3	10,783
PC-cubic 2026.05	62.1	-
Qwen2.5-32B-Instruct + Bootcamp-SFT-RL 2025.08	60.5	-
Full Attention 2026.04	60	-
Vanilla 2026.04	60	17,492
GRPO+LP 2026.04	60	11,932
GRPO-8k 2026.04	60	11,172
GRPO+SOP 2026.04	60	13,838
GRPO-λ 2026.04	60	12,814
Step-GRPO 2026.04	60	13,356
Mixture-of-Agents (MoA) 2026.04	60	-
DS-R1-Distilled-Qwen-32B + Bootcamp-RL 2025.08	56.8	-
DiffMAS 2026.04	56.7	-
GRPO-8k 2026.04	53.3	12,638
TextMAS 2026.04	53.3	-
LatentMAS 2026.04	53.3	-
OPD 2026.05	52.5	-
DS-R1-Distilled-Qwen-32B 2025.08	52.5	-
PRUNE-OPD (overlap) 2026.05	52.1	-
Upper Bound 2025.08	51.11	-
PiCSAR 2025.08	51.11	-
Standard MV 2026.05	50.8	-
DEER+SFT 2026.04	50	13,330
GSPO 2026.04	50	-
LatentMAS 2026.04	50	-
DiffMAS 2026.04	50	-
CIKA 2026.05	50	-
EVOTD 2026.05	49.9	-
TriAttention 2026.04	49.2	-
GCPO 2026.05	49	-
OPD (Truncate 4k) 2026.05	48.8	-
EVOTD 2026.05	47.6	-
EVOTD 2026.05	47.3	-
e3-1.7B 2026.04	46.7	11,804
DEER+SFT 2026.04	46.7	12,503
Single 2026.04	46.7	-
GRPO 2026.04	46.67	-
SPS 2026.04	46.67	-
Quest 2026.05	46.67	-
SnapKV 2026.05	46.67	-
DQO 2026.05	46.5	-
DIVER 2026.05	44.2	-
Evol-Instruct 2026.05	44.2	-
GRPO-λ 2026.04	43.3	12,994
Single 2026.04	43.3	-
TextMAS 2026.04	43.3	-
Div-R1 2026.05	42.5	-
HAPO 2025.09	41.71	-
Average 2025.08	41.3	-
Qwen2.5-32B-Instruct + Bootcamp-SFT 2025.08	41.2	-
Full Attention 2026.04	40.8	-
Qwen3-1.7B 2026.04	40	15,796
JustRL-deepseek-1.5B 2026.04	40	8,482
GRPO 2026.04	40	16,427
Step-GRPO 2026.04	40	13,146
DAPO 2026.04	40	-
DAPO 2026.05	40	-

Showing 100 of 311 rows