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

Knowledge Reasoning on MMLU-Pro

91.43Accuracy

SIGMA

Updated 5d ago

Evaluation Results

Method
SIGMA 2026.05	91.43	93.91	-	-	-
G-Designer 2026.05	90.54	93.22	-	-	-
GoA 2026.05	90.24	92.86	-	-	-
GPTSwarm 2026.05	90.12	92.39	-	-	-
ComplexCoT 2026.05	86.45	91.01	-	-	-
CoT 2026.05	85.72	90.68	-	-	-
In-place 2025.10	83.4	-	-	-	-
In-place 2025.10	81.7	-	-	-	-
Memory 2025.10	81.4	-	-	-	-
In-place 2025.10	81.3	-	-	-	-
MoA 2026.05	80.76	89.83	-	-	-
Multi-turn 2025.10	80.5	-	-	-	-
Memory 2025.10	80.3	-	-	-	-
Memory 2025.10	80.2	-	-	-	-
Qwen3 2026.04	80.1	-	-	-	-
I-DLM 2026.04	79.7	-	-	-	-
Multi-turn 2025.10	79.3	-	-	-	-
In-place 2025.10	79.3	-	-	-	-
In-place 2025.10	79.1	-	-	-	-
Memory 2025.10	78.7	-	-	-	-
Memory 2025.10	78.5	-	-	-	-
Multi-turn 2025.10	77.9	-	-	-	-
In-place 2025.10	77.7	-	-	-	-
Qwen2.5-32B-Instruct + Bootcamp-SFT-RL 2025.08	77.1	-	-	-	-
Multi-turn 2025.10	77.1	-	-	-	-
LLaDA-2.1-flash 2026.04	76.6	-	-	-	-
Memory 2025.10	76.5	-	-	-	-
Multi-turn 2025.10	76.3	-	-	-	-
SCOPE-RL 2025.10	76.1	-	-	-	-
Qwen2.5-32B-Instruct + Bootcamp-SFT 2025.08	76.1	-	-	-	-
Vanilla 2026.05	75.71	87.69	-	-	-
KL-cov 2025.10	75.5	-	-	-	-
Multi-turn 2025.10	75.3	-	-	-	-
Qwen3 2026.04	75.1	-	-	-	-
CISPO 2025.10	75.1	-	-	-	-
DAPO 2025.10	75	-	-	-	-
LLaDA-2.0-flash 2026.04	74.8	-	-	-	-
Entropy-Adv 2025.10	74.8	-	-	-	-
Memory 2025.10	74.6	-	-	-	-
In-place 2025.10	74.4	-	-	-	-
GRPO 2025.10	74.1	-	-	-	-
Entropy-Reg 2025.10	73.7	-	-	-	-
Multi-turn 2025.10	73.4	-	-	-	-
I-DLM 2026.04	73.1	-	-	-	-
DS-R1-Distilled-Qwen-32B 2025.08	73	-	-	-	-
Base 2025.10	72.6	-	-	-	-
DS-R1-Distilled-Qwen-32B + Bootcamp-RL 2025.08	70.6	-	-	-	-
In-place 2025.10	70.6	-	-	-	-
Multi-turn 2025.10	70.2	-	-	-	-
Qwen2.5-32B-Instruct 2025.08	69.1	-	-	-	-
Memory 2025.10	68.8	-	-	-	-
Qwen2.5-32B-Instruct + Bootcamp-RL 2025.08	68.3	-	-	-	-
LLaDA-2.1-mini 2026.04	64.8	-	-	-	-
SDPO 2026.02	63.3	-	-	-	-
Qwen3-8B 2026.02	62.5	-	-	-	-
SDAR 2026.04	61.5	-	-	-	-
Multi-turn 2025.10	61.3	-	-	-	-
Memory 2025.10	61.3	-	-	-	-
In-place 2025.10	61.3	-	-	-	-
LOVER (sum) 2026.05	61.24	-	-	-	-
Supervised 2026.05	61.13	-	-	-	-
SCOPE-RL 2025.10	60.7	-	-	-	-
DAPO 2025.10	59.8	-	-	-	-
LOVER (max) 2026.05	59.28	-	-	-	-
KL-cov 2025.10	59.1	-	-	-	-
CoT-Decoding (sum) 2026.05	58.81	-	-	-	-
GRPO 2025.10	58.2	-	-	-	-
Entropy-Reg 2025.10	58.2	-	-	-	-
Qwen3-4B 2026.02	58.1	-	-	-	-
SDPO 2026.02	58	-	-	-	-
CISPO 2025.10	57.8	-	-	-	-
Entropy-Adv 2025.10	57.7	-	-	-	-
Majority Voting 2026.05	57	-	-	-	-
SDAR 2026.04	56.9	-	-	-	-
Supervised 2026.05	55.86	-	-	-	-
LOVER (sum) 2026.05	55.61	-	-	-	-
Greedy 2026.05	54.93	-	-	-	-
Multi-turn 2025.10	54.8	-	-	-	-
Memory 2025.10	54.8	-	-	-	-
In-place 2025.10	54.8	-	-	-	-
CoT-Decoding (sum) 2026.05	54.61	-	-	-	-
CoT-Decoding (max) 2026.05	54.22	-	-	-	-
LOVER (max) 2026.05	52.65	-	-	-	-
Majority Voting 2026.05	50.58	-	-	-	-
Base 2025.10	50	-	-	-	-
Dream 7B-Instruct (Top Margin + Suffix-Anchored Confidence Modulation) 2026.05	49.4	-	-	-	-
Dream 7B-Instruct (Top Probability + Suffix-Anchored Confidence Modulation) 2026.05	48.83	-	-	-	-
DAPO 2025.10	48.6	-	-	-	-
Dream 7B-Instruct (Top Margin + Suffix Anchor) 2026.05	48.39	-	-	-	-
KL-cov 2025.10	48	-	-	-	-
SCOPE-RL 2025.10	48	-	-	-	-
Dream 7B-Instruct (Top Probability + Suffix Anchor) 2026.05	47.85	-	-	-	-
Entropy-Adv 2025.10	47.3	-	-	-	-
CoT-Decoding (max) 2026.05	45.92	-	-	-	-
Greedy 2026.05	45.67	-	-	-	-
GRPO 2025.10	44.8	-	-	-	-
CISPO 2025.10	44.1	-	-	-	-
Dream 7B-Instruct (Top Margin) 2026.05	43.38	-	-	-	-
Entropy-Reg 2025.10	42.7	-	-	-	-
Supervised 2026.05	41.93	-	-	-	-

Showing 100 of 218 rows