NEWHow AI models draw the world map · Gecko Tests
Beta
Home/Models/Gemma 4 31B
Google DeepMind logo

Gemma 4 31B

by Google DeepMind · Released Apr 2026

Open SourceMultimodal
53.3
avg score
Rank #126
Compare
Better than 60% of all models
Context
262K tokens (~131 books)
Input $/1M
$0.09
Output $/1M
$0.34
Type
multimodal
License
Open Source
Benchmarks
29 tested
Data as of
About

Gemma 4 31B Instruct is Google DeepMind's 30.7B dense multimodal model supporting text and image input with text output. Features a 256K token context window, configurable thinking/reasoning mode, native function...

Tested on 29 benchmarks · BenchGecko score 53.3. Top scores: Chatbot Arena Elo — Overall (1452.8%), Chatbot Arena Elo — Coding (1364.7%), Artificial Analysis · GPQA Diamond (85.7%).

Capabilities
coding
50.9
#89 globally
reasoning
59.1
#60 globally
math
73.6
#36 globally
knowledge
34.9
#227 globally
speed
44.7
#44 globally
general
49.3
#33 globally
language
69.5
#80 globally
Benchmark Scores
Compare All
Tested on 29 benchmarks · Ranked across 8 categories
Score Distribution (all 312 models)
0255075100
▲ You are here
codingCompare coding →
LiveBench — Coding

Regularly refreshed coding problems that avoid data contamination. New problems added monthly to prevent memorization.

60.3·
WeirdML

Unusual and adversarial machine learning challenges. Tests robustness of reasoning about edge cases in ML systems.

52.3·
LiveBench — Agentic Coding

LiveBench coding tasks that require multi-step reasoning and tool use. Tests planning and execution of complex coding workflows.

40.0·
reasoningCompare reasoning →
LiveBench — Reasoning

Regularly refreshed reasoning problems testing logical deduction, spatial reasoning, and analytical thinking.

59.4·
LiveBench — Data Analysis

Fresh data analysis tasks testing ability to interpret tables, charts, and statistical data.

58.8·
mathCompare math →
LiveBench — Mathematics

Regularly updated math problems that test numerical reasoning, algebra, calculus, and combinatorics.

73.9·
OTIS Mock AIME 2024-2025

Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.

73.3·
Quick compare:meituan logovs LongCat Flash ChatDeepSeek logovs DeepSeek R1 Distill Llama 8BxAI logovs Grok 3 Mini Beta
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Recently Happened
Gemma 4 31B pricing dropped 10%
Aug 26, 2026
Gemma 4 31B pricing increased 11%
Aug 21, 2026
Gemma 4 31B pricing dropped 10%
Aug 19, 2026
Gemma 4 31B pricing dropped 7%
Apr 14, 2026
Gemma 4 31B added
Apr 5, 2026
Similar Models
meituan logo
LongCat Flash Chat
meituan
53.4$0.20/1M
DeepSeek logo
DeepSeek R1 Distill Llama 8B
DeepSeek
53.1n/a
xAI logo
Grok 3 Mini Beta
xAI
53.1$0.30/1M
Links
Info
Google DeepMindPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@Google DeepMind
Source Code
GitHubHugging Face
BenchGecko API
gemma-4-31b-it
Specifications
  • Typemultimodal
  • Context262K tokens (~131 books)
  • ReleasedApr 2026
  • LicenseOpen Source
  • StatusActive
  • Cost / Message~$0.001
Available On
Google DeepMind logoGoogle DeepMind$0.09
Categories
codingreasoningmathknowledgespeedgenerallanguage
Learn More
context-windowtransformeropen weightstokens
Share & Export
Tweet
Related Models
meituan logoLongCat Flash Chat
53.4
DeepSeek logoDeepSeek R1 Distill Llama 8B
53.1
xAI logoGrok 3 Mini Beta
53.1
xAI logoGrok 4 Fast
53.1
Gemma 4 31B is an open-source multimodal AI model by Google DeepMind, released in April 2026. It has an average benchmark score of 53.3. Context window: 262K tokens.

Related Models

LongCat Flash Chat · meituanDeepSeek R1 Distill Llama 8B · DeepSeekGrok 3 Mini Beta · xAIGrok 4 Fast · xAIQwen3.5-35B-A3B · Alibaba Qwen

Benchmarks

Chatbot Arena Elo — OverallChatbot Arena Elo — CodingArtificial Analysis · GPQA DiamondArtificial Analysis · IFBenchLiveBench — Mathematics

Related Pages

Google DeepMind · ProviderGoogle DeepMind · EconomyAll ModelsCompare ModelsPricingDevelopers · APIContext Window · Glossary

Key facts · as of 2026-10-05

  • Gemma 4 31B by Google DeepMind. BenchGecko score 53.3, rank 126 of 312 scored models (normalized average of public benchmark scores).
  • List price $0.0900 input · $0.34 output per 1M tokens (as of 2026-10-05).
  • Sold by 13 providers (as of 2026-10-05): DeepInfra (fp4) $0.0900 in / $0.34 out · CoreWeave (fp4) $0.10 in / $0.34 out · Chutes (fp4) $0.12 in / $0.37 out · Venice (fp4) $0.12 in / $0.36 out · Crusoe (bf16) $0.14 in / $0.40 out · and 8 more. Every provider

How to cite · data as of 2026-10-05

Gemma 4 31B · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/gemma-4-31b-it

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP