Translate
Independent model intelligence

Find the right AI model,
not just the loudest one.

Compare current model capability, context, modalities, public preference and API economics in one continuously refreshed decision surface.

LiveBench 2.0 LMArena models.dev OpenRouter AIHOT cross-check
Back to leaderboard
Google · Active

Gemma 4 31B

gemma family · launched 2026-04-02

MultimodalReasoningToolsStructuredOpen weights
AITopic score48.4Not yet ranked

Gemma 4 31B Dense is designed for advanced reasoning, agentic workflows, and longer context and is natively trained on 140+ languages.

Evidence coverage95%5 evaluations · 2 institutions
Context / maximum output262.1k / 262.1ktokens
Cache / input / output$0.1 · $0.1 · $0.34USD per million tokens
Value / freshness67.9 / 61relative value · release freshness
Product profile

What this model supports

Inputtext, image
Outputtext
Knowledge cutoff2025-01
WeightsOpen weights
Capability profile

Strengths by domain

Coding49.01 public results
Research & reasoning68.31 public results
Professional work30.61 public results
Knowledge & instruction44.41 public results
Behind the score

Every result has a source

Raw results and standardized readings are shown together. Missing tests add neither points nor penalties.

Coding1 results

LiveBench 2.0 · CodingLiveBench49.00 scoreReading 49.0

Imported from LiveBench 2.0. The published 0–100 score is used directly; it is not z-score normalized.

Configuration: LiveBench 2.0Open evidence

Research & reasoning1 results

LiveBench 2.0 · ResearchLiveBench68.30 scoreReading 68.3

Imported from LiveBench 2.0. The published 0–100 score is used directly; it is not z-score normalized.

Configuration: LiveBench 2.0Open evidence

Professional work1 results

LiveBench 2.0 · ProfessionalLiveBench30.60 scoreReading 30.6

Imported from LiveBench 2.0. The published 0–100 score is used directly; it is not z-score normalized. Professional = long-agentic, browser, data-analysis and design average.

Configuration: LiveBench 2.0Open evidence

Knowledge & instruction1 results

LiveBench 2.0 · KnowledgeLiveBench44.40 scoreReading 44.4

Imported from LiveBench 2.0. The published 0–100 score is used directly; it is not z-score normalized. Knowledge uses the published LiveBench research suite.

Configuration: LiveBench 2.0Open evidence

Human preference & indices3 results

AA IndexArtificial Analysis64.02 indexReading 55.4

Imported from Artificial Analysis. The published 0–100 score is used directly; it is not z-score normalized.

Configuration: LiveBench 2.0Open evidence
Arena TextLMArena1,443.0 EloReading 48.4

Direct LMArena text leaderboard result.

Open evidence
LiveBenchLiveBench40.59 scoreReading 40.6

Imported from LiveBench 2.0. The published 0–100 score is used directly; it is not z-score normalized.

Configuration: LiveBench 2.0Open evidence
Same builder

More from Google