UnifyBench

Weighted consensus · Public benchmarks

LLM reasoning leaderboard

Compare AI reasoning models and measured effort settings across published reasoning benchmarks.

Looking for which LLM is best? UnifyBench is a first stop: source-balanced consensus across many published benchmarks (not one board). Shortlist here, then validate on Artificial Analysis, check Arena for preference, and finish with task-specific evals.

See the full ranking

Reasoning leaders

Comparison score · 0–100

Experimental scores for the selected profile and sources. Small gaps may not be meaningful. How scoring works · Effort evidence collected

471unique models568model configurations134source pages
Explore all 134 source pages Independent evaluations and lab reports

Each model is counted once across effort levels. Configurations count documented model + effort settings, including unranked.

Original reports and benchmark leaderboards. Distinct source pages with published results in the catalog. Multiple results and page sections count as one source.

One entry per model, using its highest-scoring eligible effort setting. How effort ranking works.

Customize ranking Sources, profiles and evidence details · Including unranked entries
All presets

Evidence only MultimodalLong contextHuman prefAll capabilitiesThese profiles do not yet have enough comparable evidence for a ranking.

Adjust weights

Full ranking

471 entries
Compare two configurations

Custom capability ranking of measured model configurations. Scores are experimental; missing results remain unknown.
RankModelScoreCapability coverage
—MiniMax-Text-01 · Effort unspecifiedMiniMax—Not scoredProvisional · 0%0 families · 0 selected capabilities
—MiniMax-VL-01 · Effort unspecifiedMiniMax—Not scoredNo evidence0 families · 0 selected capabilities
—Ministral 3B · UnspecifiedMistral—Not scoredProvisional · 0%0 families · 0 selected capabilities
—Ministral 8B · UnspecifiedMistral—Not scoredProvisional · 0%0 families · 0 selected capabilities
—Mistral Large 2.1 · UnspecifiedMistral—Not scoredProvisional · 0%0 families · 0 selected capabilities
—Mistral Nemo 12B · UnspecifiedMistral—Not scoredProvisional · 0%0 families · 0 selected capabilities
—Mistral Small Creative · Effort unspecifiedMistral—Not scoredNo evidence0 families · 0 selected capabilities
—Muse Glimmer 30B · XHighMeta—Not scoredProvisional · 0%0 families · 0 selected capabilities
—North-Micro-Vision-Instruct · Effort unspecifiedCohere—Not scoredNo evidence0 families · 0 selected capabilities
—North-Mini-Code-1.0 · Effort unspecifiedCohere—Not scoredNo evidence0 families · 0 selected capabilities
—NVIDIA-Nemotron-Nano-12B-v2 · Effort unspecifiedNVIDIA—Not scoredNo evidence0 families · 0 selected capabilities
—NVIDIA-Nemotron-Nano-9B-v2-Japanese · Effort unspecifiedNVIDIA—Not scoredNo evidence0 families · 0 selected capabilities
—o1-pro · Effort unspecifiedOpenAI—Not scoredNo evidence0 families · 0 selected capabilities
—Phi-3-medium-128k-instruct · UnspecifiedMicrosoft—Not scoredProvisional · 0%0 families · 0 selected capabilities
—Phi-3-medium-4k-instruct · Effort unspecifiedMicrosoft—Not scoredProvisional · 0%0 families · 0 selected capabilities
—Phi-3-mini-128k-instruct · Effort unspecifiedMicrosoft—Not scoredProvisional · 0%0 families · 0 selected capabilities

More models on this page

About this ranking

This profile gives hard reasoning all the weight. It compares identified configurations using their own measured results, rather than borrowing results from another effort setting. Rankings require support in this capability. Matching effort labels across providers do not imply equal compute budgets.

View the Capability ranking →

Behind the ranking

294 entries with published results · 233 ranked · 0 provisional estimates.

Explore seven capabilities, the original results, and gaps in the evidence.

Coding rankingReasoning rankingAgentic rankingKnowledge rankingCompare capabilities ↗Published benchmarks ↗Evidence coverage ↗
Cite this ranking

Include the model and effort setting, the selected profile, and the dated data downloads. A shared link uses the latest data; it does not freeze a ranking in time.

UnifyBench ranking
Data updated: 22 Oct 2026
Mode: custom; evidence: documented; configurations: best
Weights: v3;agentic:0,hard_reasoning:100,coding:0,human_pref:0,knowledge:0,multimodal:0,long_context:0
https://unifybench.ai/rankings/reasoning?evidence=documented&page=16&view=overview

Catalog data · Effort measurements · Configuration scoring rules · Methodology

Catalog dated 22 Oct 2026. Effort evidence collected 18 Oct 2026. Collection dates are not evaluation dates.

What does the score mean?

The adjusted comparison score estimates performance against the same reference panel from matched published comparisons. Capability pools matched comparisons across families and checks aggregate breadth and independent source coverage. Custom profiles require support in every selected capability. Missing results never count as losses.

Family details show observed win shares against available reference peers. Those descriptive values are supporting evidence, not adjusted capability scores. Provider reports and independent results remain labeled. Matching reported settings do not establish equal inference effort, and score differences do not establish statistical significance.

Read the full methodology →