AI Paper Atlas

Research map / Benchmarks / MMLU-Pro

MMLU-Pro leaderboard

73 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on MMLU-Pro. Best accuracy: 86.25 (Greedy LinUCB, AAAI 2026).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

MMLU-Pro · accuracy

58 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
186.25Greedy LinUCB
6-LLM pool
Online Multi-LLM Selection via Contextual Bandits Under Unstructured Context Evolution
AAAI 2026 · Table 1
Contextual bandits and Thompson sampling
286Team (Expert Not Mentioned)
heterogeneous 4-model team
Multi-Agent Teams Hold Experts Back
ICML 2026 · Table 2
LLM agents
382.06MXFP4-MBS-H
DeepSeek-R1
Unveiling the Potential of Quantization with MXFP4: Strategies for Quantization Error Reduction
ICML 2026 · Table 3
Quantization
481.8LASE
GPT-4.1-mini
Efficient Multi-Agent Reasoning via Confidence-Guided Adaptive Debate
ICML 2026 · Table 1
LLM agents
581.3RRM-32B (knockout best-of-32)
DeepSeek-R1-Distill-Qwen-32B
Reward Reasoning Models
NeurIPS 2025 · Reward-guided best-of-N table
Reasoning and chain of thought
679.43Adaptive Think (α=0.1)
DeepSeek-R1-32B
Think or Not? Exploring Thinking Efficiency in Large Reasoning Models via an Information-Theoretic Lens
NeurIPS 2025 · Table 2
Reasoning and chain of thought
779.4Full reasoning trace (decile 100)
gpt-oss-120b
Probing the Trajectories of Reasoning Traces in Large Language Models
NeurIPS 2026 · Table A2
Reasoning and chain of thought
878.63Star Elastic 30B (3.6A)
Nemotron Nano v3 30B
Star Elastic: Many-in-One Reasoning LLMs with Efficient Budget Control
ICML 2026 · Table 1
Reasoning and chain of thought
978Targeted alternation training
GPT-4.1-mini
Learning Rewrite-Invariant Reasoning with Targeted Alternation Training
ICML 2026 · Table 1
Instruction and fine-tuning
1077.39FLARE-9B (AR-Trust)
Qwen3.5-9B
FLARE: Diffusion for Hybrid Language Model
NeurIPS 2026 · Table 1
Language, safety and interpretability
48 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

MMLU-Pro · score

15 papers, each with its best reported number for its own method; higher is better.

#score ↑MethodPaperTopic
186.78JiSi (Ours)Beyond Gemini-3-Pro: Revisiting LLM Routing and Aggregation at Scale
ICML 2026 · Table 1
Agents, code and math
278DR-Rubric-8B (BS-3)
Qwen3-8B-SFT
Deep Research as Rubric
NeurIPS 2026 · Table 2
Reasoning and chain of thought
372.25Ultra-253B-Base
Llama-3.1-405B-Instruct
FFN Fusion: Rethinking Sequential Computation in Large Language Models
NeurIPS 2025 · Table: Comparison of Ultra-253
Attention and transformers
467.9OVR-7B
Qwen2.5-VL-7B
Open Vision Reasoner: Transferring Linguistic Cognitive Behavior for Visual Reasoning
NeurIPS 2025 · Language benchmark table
Multimodal LLMs
566.21Avengers
10 open-source ~7B models
The Avengers: A Routing Recipe for Collective Intelligence in Language Models
AAAI 2026 · Table 2
Language, safety and interpretability
647.9Dream-Instruct + G-Star
Dream-Instruct 7B
Guided Star-Shaped Masked Diffusion
ICML 2026 · Table 1
Fast diffusion sampling
747.6UniGist
Llama-3.1-8B-Instruct
UniGist: Towards General and Hardware-aligned Sequence-level Long Context Compression
NeurIPS 2025 · Short-context table
Attention and transformers
846.2Llama-3.1-8B-Instruct + Paprika
Llama-3.1-8B-Instruct
Training a Generally Curious Agent
ICML 2025 · Table: Evaluation of Paprika o
Human-AI and goal-conditioned
943.7Webscale-RL
Qwen2.5-3B
Webscale-RL: Automated Data Pipeline for Scaling RL Data to Pretraining Levels
ICLR 2026 · Table 1
Reasoning and chain of thought
1036.68SimPER
Llama3-8B-Instruct
SimPER: A Minimalist Approach to Preference Alignment without Hyperparameters
ICLR 2025 · Open LLM Leaderboard results t
Alignment and preferences
5 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.