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

Reasoning on MMLU-Pro

92.86Accuracy

Agent Q-Mix

Updated 1mo ago

Evaluation Results

Method
Agent Q-Mix 2026.04	92.86	-
Lobster 2026.04	91.43	-
LangGraph 2026.04	91.43	-
Gemini-3.0 Pro 2026.01	90.1	-
Agent Framework 2026.04	90	-
TopoDIM 2026.04	88.57	-
AutoGen 2026.04	88.57	-
Agent Q-Mix 2026.04	88.57	-
Claude Sonnet 4.5 2026.01	88.2	-
GPT-5 High 2026.01	87.5	-
Lobster 2026.04	87.14	-
GTD 2026.04	87.14	-
Qwen3.5-27B 2026.04	86.1	-
GTD 2026.04	85.71	-
MaAS 2026.04	85.71	-
TopoDIM 2026.04	85.71	-
In-place 2025.10	85.4	-
DeepSeek-V3.2 Thinking 2026.01	85	-
In-place 2025.10	85	-
MiMo-V2-Flash 2026.01	84.9	-
In-place 2025.10	84.8	-
Kimi-K2 Thinking 2026.01	84.6	-
In-place 2025.10	84.6	-
G-Designer 2026.04	84.29	-
MaAS 2026.04	84.29	-
In-place 2025.10	84.2	-
K-EXAONE-236B-A23B 2026.04	83.8	-
Qwen3-VL-235B-A22B 2026.04	83.8	-
In-place 2025.10	83.8	-
EXAONE 4.5 33B 2026.04	83.3	-
GPT-5 mini 2026.04	83.3	-
In-place 2025.10	83	-
AgentDropout 2026.04	82.86	-
LLM-Debate 2026.04	82.86	-
GPTSwarm 2026.04	82.86	-
In-place 2025.10	82.8	-
SMCS 2025.07	82.02	-
In-place 2025.10	81.9	-
GPTSwarm 2026.04	81.43	-
AutoGen 2026.04	81.43	-
Base (direct) 2026.04	81.43	-
QwQ-32B 2025.09	81.4	-
QwQ-32B-CodeGym 2025.09	81.4	-
In-place 2025.10	81.3	-
In-place 2025.10	81.3	-
Memory 2025.10	80.9	-
Majority Voting 2025.07	80.85	-
Memory 2025.10	80.7	-
In-place 2025.10	80.7	-
Symbolic-MoE* 2025.07	80.6	-
Multi-turn 2025.10	80.5	-
GPT-4.1 2025.07	80.43	-
Multi-turn 2025.10	80.1	-
In-place 2025.10	80.1	-
Agent Framework 2026.04	80	-
Memory 2025.10	79.7	-
In-place 2025.10	79.7	-
MoA 2025.07	79.6	-
In-place 2025.10	79.5	-
Multi-turn 2025.10	79.3	-
Multi-turn 2025.10	79.1	-
Memory 2025.10	78.7	-
LLM-Debate 2026.04	78.57	-
G-Designer 2026.04	78.57	-
Memory 2025.10	78.5	-
In-place 2025.10	78.5	-
Claude-3.5-Sonnet 2025.07	78.34	-
Memory 2025.10	78.3	-
FP16 2025.11	78.1	-
Multi-turn 2025.10	78.1	-
Self Consistency (Best on Validation) 2025.07	78.01	-
QTIP 2025.11	77.9	-
Qwen3-32B 2025.07	77.76	-
Open-source Upper Bound 2025.07	77.76	-
Memory 2025.10	77.7	-
In-place 2025.10	77.7	-
ParoQuant 2025.11	77.5	-
Multi-turn 2025.10	77.3	-
Memory 2025.10	77.3	-
AWQ 2025.11	77.2	-
GLM-Z1-32B-0414 2025.07	76.76	-
E-QAT 2025.11	76.5	-
In-place 2025.10	76.5	-
In-place 2025.10	76.1	-
Memory 2025.10	75.9	-
Memory 2025.10	75.7	-
DeepSeek-R1-Distill-Qwen-32B 2025.07	75.17	-
Multi-turn 2025.10	75.1	-
Multi-turn 2025.10	75.1	-
Multi-turn 2025.10	74.8	-
DeepSeek-R1-Distill-Llama-70B 2025.07	74.75	-
QwQ-32B 2025.07	74.67	-
FP16 2025.11	74.6	-
Base (direct) 2026.04	74.29	-
AgentDropout 2026.04	74.29	-
In-place 2025.10	74.2	-
HuatuoGPT-o1-72B 2025.07	74.16	-
ParoQuant 2025.11	74.1	-
QTIP 2025.11	74	-
GPT-o3-mini 2025.07	74	-

Showing 100 of 311 rows