# Google Gemma 4 31B As of October 22, 2026, Google Gemma 4 31B ranks #89 of 123 models overall with 42.1 ±14.7 on the LLMBenchmark Index. Gemma 4 31B ranks #50 of 108 in math, its strongest category. Google Gemma 4 31B costs $0.14 per million input tokens and $0.40 per million output tokens. Gemma 4 31B was released on April 2, 2026. Page: https://llmbenchmark.ai/models/gemma-4-31b-it Data from October 22, 2026. Every figure below names the board it was read from and the day. - Maker: Google - Released: April 2, 2026 - Price per million input tokens: $0.14 - Price per million output tokens: $0.40 - Context window: 262K tokens - Overall rank: 89 of 123, by the LLMBenchmark Index - LLMBenchmark Index: 42.1 ±14.7 (90% margin, from 7 of 7 inputs) - Artificial Analysis Intelligence Index: 14.7 - Vibe Coding Index: 39.0 (from 2 of 6 coding benchmarks) Benchmark results (71 of 195): - LiveBench coding: 49.0% (source: LiveBench, read October 22, 2026) - LMArena WebDev: 1367 Elo (source: LMArena WebDev, read October 22, 2026) - AA Intelligence Index: 14.7 (source: Artificial Analysis (via OpenRouter), read October 22, 2026) - AA Coding Index: 43.4 (source: Artificial Analysis (via OpenRouter), read October 22, 2026) - AA Agentic Index: 4.2 (source: Artificial Analysis (via OpenRouter), read October 22, 2026) - LiveBench reasoning: 68.3% (source: LiveBench, read October 22, 2026) - LiveBench data analysis: 53.3% (source: LiveBench, read October 22, 2026) - LiveBench research: 44.4% (source: LiveBench, read October 22, 2026) - LiveBench browser: 4.0% (source: LiveBench, read October 22, 2026) - LiveBench long agentic: 13.3% (source: LiveBench, read October 22, 2026) - LiveBench design: 58.4% (source: LiveBench, read October 22, 2026) - SAGE (Vals AI): 55.0% (source: Vals AI boards, read October 22, 2026) - CaseLaw (Vals AI): 52.6% (source: Vals AI boards, read October 22, 2026) - MortgageTax (Vals AI): 61.4% (source: Vals AI boards, read October 22, 2026) - GPQA Diamond (Artificial Analysis run): 85.7% (source: Artificial Analysis, read October 22, 2026) - Humanity's Last Exam (Artificial Analysis run): 23.6% (source: Artificial Analysis, read October 22, 2026) - SciCode: 45.5% (source: Artificial Analysis, read October 22, 2026) - Terminal-Bench Hard (Artificial Analysis run): 36.4% (source: Artificial Analysis, read October 22, 2026) - IFBench: 75.6% (source: Artificial Analysis, read October 22, 2026) - Long context reasoning (AA-LCR): 69.7% (source: Artificial Analysis, read October 22, 2026) - τ²-Bench Telecom: 65.5% (source: Artificial Analysis, read October 22, 2026) - WeirdML: 52.3% (source: Epoch AI benchmarking hub, read October 22, 2026) - GPQA Diamond (Epoch AI run): 75.8% (source: Epoch AI benchmarking hub, read October 22, 2026) - OTIS Mock AIME: 73.3% (source: Epoch AI benchmarking hub, read October 22, 2026) - SimpleQA Verified: 10.4% (source: Epoch AI benchmarking hub, read October 22, 2026) - DTBench: 82.7% (source: Epoch AI benchmarking hub, read October 22, 2026) - LMCA: 39.3% (source: Epoch AI benchmarking hub, read October 22, 2026) - Surface Evolver Bench: 30.6% (source: Epoch AI benchmarking hub, read October 22, 2026) - ALE-Bench: 925.5 (source: ALE-Bench (Sakana AI), read October 22, 2026) - Epoch Capabilities Index: 143.0 (source: Epoch AI benchmarking hub, read October 22, 2026) - LMArena Text: 1443 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: coding: 1459 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: hard prompts: 1448 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: math: 1463 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: instruction following: 1433 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: creative writing: 1417 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Vision: 1276 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: English: 1451 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: non-English: 1430 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: Chinese: 1475 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: French: 1436 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: Spanish: 1442 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: Russian: 1460 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: Polish: 1433 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: multi-turn: 1452 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: longer queries: 1446 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: expert: 1464 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: hard prompts in English: 1456 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: software and IT: 1463 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: mathematical work: 1465 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: science: 1457 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: medicine and healthcare: 1451 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: legal and government: 1438 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: business and finance: 1430 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: writing and language: 1432 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Text: entertainment and media: 1405 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena WebDev: React: 1359 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena WebDev: HTML: 1380 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Vision: English: 1271 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Vision: Chinese: 1326 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Vision: reading text: 1282 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Vision: diagrams: 1283 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Vision: homework: 1310 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Vision: humor: 1280 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Vision: captioning: 1218 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Vision: recognizing things: 1226 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Vision: creative writing: 1285 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - LMArena Document: 1425 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026) - Kagi LLM Benchmark: 63.5% (source: Kagi LLM Benchmarking Project, read October 22, 2026) - SciCode: 8.9% (source: Kaggle Benchmarks, read October 22, 2026) - FACTS Benchmark Suite: 45.3% (source: Kaggle Benchmarks, read October 22, 2026)