Skip to content

Google · Chat

Gemma 4 31B IT

Google 🇺🇸ChatGemini familyOpen weights. One of its sources says the weights are publicly available.The maker lists this model as deprecated (15 Sep 2026).
From Venice's listing:
Gemma 4 31B is a dense model from Google DeepMind with 31B parameters, delivering frontier-level reasoning performance. It handles text, image, and video input, supports 256K context, function calling, and configurable thinking modes.
llminfo IQ
128.3
Maker's price
$0.18 blended / 1M
Max context
262K
Listings
170

Evidence and sources

🤗 google/gemma-4-31b-it Model architecture, as its sources state it llminfo IQ 128.3, ranked among chat models. From 333 benchmarks. Plenty of data behind it.How llminfo IQ is computedCoding IQ 101.7, 95% interval 94–109. From 13 benchmark families across 10 sources. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.How Coding IQ is computedGood value (61.9). Cheaper per task than its benchmark results predict. From 11 benchmarks with measured task costs. Plenty of data behind it.How Value for money worksYap score: tokens used per coding task vs. models with the same resultsTerse (43 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.Median 10,710 output tokens per task at default effort.How the Yap score worksReleased 2 Apr 2026Maker HQ: US (according to Google's own website)
All rankings: #28 on NanoGPT recently #45 on OpenRouter today #45 on OpenRouter this month #51 on OpenRouter this week #175 on NanoGPT recentlySources: Google Generative AI API data docs, Berget AI data, Infercom data, flex.ai API data, LiteLLM data, Models.dev data, Venice.ai data, llminfo researchPrimary data from Google Generative AI API from Berget AI: variant, primary mode from Infercom: architecture from flex.ai API: hugging face id from LiteLLM: deprecation date from Models.dev: weights availability from Venice.ai: description from llminfo research: release date
Tags from its sources:textvisioninstructtoolsreasoning

Advantages and trade-offs

Compared with 10 chat models, picked mainly by popularity and IQ

Trade-offs

Deprecated by its maker

The maker has already deprecated this model (15 Sep 2026).

We only make a claim when there's enough data on models of the same type. Updated 22 Nov, 04:34 UTC.Prices come from the model's maker where they sell it, otherwise their official partner, otherwise the market median. This comparison uses 9 maker's prices; 1 market median price.

Input & output

as stated by the model's providers
Text
Images
Video
Files
Text

Capabilities

Supported operations and provider-stated features

VisionReasoningFunction callingStructured outputPrompt cachingWeb searchBatch API

Listings

 Googlemodels/gemma-4-31b-itSources: GenAI Prices data, Google Generative AI API data docs, LiteLLM data, Models.dev data262K$0.12cache read $0.09$0.36——
 Aihubmixgemma-4-31b-it-free · freeSource: AIHubMix data docs262K$0.00$0.00——
 Flowdotgoogle/gemma-4-31b-it:free · freeSource: FlowDot data262K$0.00$0.00——
 OpenRouterInference by Googlegoogle/gemma-4-31b-it:free · freeSource: OpenRouter data docs262K$0.00$0.00251,391ms
 Bothubgemma-4-31b-it:free · freeSource: Models.dev data262K$0.00$0.00——
 Unoroutergemma-4-31b-it:free · freeSource: Models.dev data262K$0.00$0.00——
 NanoGPTInference by Sail ResearchRegions: GLBgoogle/gemma-4-31b-it · standard · fp4Source: NanoGPTNot published by the provider$0.0735cache read $0.0525$0.4230.73,170ms
 Adaptiongemma-4-31b-itSource: Models.dev data256K$0.09cache read $0.05$0.34——
 Fastroutergoogle/gemma-4-31b-itSources: FastRouter data, Models.dev data262K$0.10$0.30——
 Hugging FaceInference by Coreweavegoogle/gemma-4-31B-itSource: Hugging Face Inference Providers data262K$0.10$0.34——

14 listings without a price are hidden.

Benchmarks

Named effort levels run from lowest to highest. Select a result to see its source details.

Multilingual

Benchmark Nusantara20 related benchmarks · 22 results
Show 16 more resultsHide 16 results
jv-B arti-kata (prompt arti-kata-v1, scorer kunci-alias-v1, items 4e682041d284, key 873b6f168502, public slice)
Source: Benchmark Nusantara (ibahasa) data docs
repetition: 1, route google/gemma-4-31b-it, token cap unrecorded
repetition: 2, route google/gemma-4-31b-it, token cap 600
repetition: 3, route google/gemma-4-31b-it, token cap 600
EKI EstonianSource: Keelemudelite mõõdupuu data docs
IdrockBench Uzbek4 related benchmarks · 4 results
MMMLU2 results
Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 Technical Report · Table 5 · thinking mode unless noted
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
NepNewsBench v2 draft40 related benchmarks · 40 results
entity F1 (all questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
core fact recall (all questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
core fact recall (English-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
core fact recall (Nepali-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
coverage precision proxy (all questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
Show 34 more resultsHide 34 results
effective score (all questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
effective score (English-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
effective score (Nepali-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
English prose quality (all questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
entity F1 (English-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
entity F1 (Nepali-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
entity precision (all questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
entity precision (English-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
entity precision (Nepali-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
entity recall (all questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
entity recall (English-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
entity recall (Nepali-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
Nepali prose quality (all questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
supporting fact recall (all questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
supporting fact recall (Nepali-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
valid-answer quality (all questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off
valid-answer quality (Nepali-source questions)
Source: NepNewsBench v2 (draft candidate release) data · Sep 2026
Venice via openrouter, reasoning off

Human preference

Arena ELO44 related benchmarks · 44 results
LMArena EloGemma 4 Technical Report · Table 4 · blind human preference · leaderboard snapshot 2026-06-19 · Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.

Vision

Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 Technical Report · Table 12 · thinking · 280 vision tokens
Gemma 4 Technical Report · Table 6 · thinking · 1120 vision tokens
MMMU-Pro3 results
Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 Technical Report · Table 12 · thinking · 280 vision tokens
Gemma 4 Technical Report · Table 6 · thinking · 1120 vision tokens
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated

Science

GPQASource: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
GPQA Diamond24 results
Source: Epoch AI Benchmarking Hub data
Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 Technical Report · Table 5 · thinking mode unless noted
Source: OpenRouter benchmark exports
2 Apr 2026 version · 90-day results
Across 21 providers lowest 37.9% median 84.4% highest 86.1%OpenRouter's automatic routing 82.5%
Show results from 21 providers and OpenRouter's automatic routingHide results from 21 providers and OpenRouter's automatic routing
  • DekaLLM
  • SiliconFlow
  • Crusoe
  • Parasail
  • Together
  • NextBit
  • ModelRun
  • Chutes
  • KT
  • Novita
  • Modal
  • PrimeIntellect
  • SambaNova
  • Phala
  • DeepInfra
  • CoreWeave
  • OpenRouter's automatic routing
  • Cerebras
  • Friendli
  • OpenInference
  • Venice
  • Io Net

Coding

Arena ELOSource: LMArena data · Nov 2026
Codeforces RatingGemma 4 Technical Report · Table 5 · thinking mode unless noted · Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 Technical Report · Table 5 · thinking mode unless noted
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
SWE-rebenchSource: SWE-rebench

Agentic

Source: OpenRouter benchmark exports
2 Apr 2026 version · 90-day results
Across 19 providers lowest 42.3% median 73.8% highest 77.1%OpenRouter's automatic routing 75.3%
Show results from 19 providers and OpenRouter's automatic routingHide results from 19 providers and OpenRouter's automatic routing
  • DeepInfra
  • PrimeIntellect
  • SambaNova
  • Friendli
  • Venice
  • OpenRouter's automatic routing
  • ModelRun
  • OpenInference
  • KT
  • Phala
  • CoreWeave
  • NextBit
  • DekaLLM
  • Novita
  • SiliconFlow
  • Crusoe
  • Cerebras
  • Parasail
  • Chutes
  • Modal

Reasoning

Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 official model card · no tools
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated
t2-benchSource: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim

General

Instruction following

IFBenchGemma 4 Technical Report · Table 5 · thinking mode unless noted · Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
IFEvalGemma 4 Technical Report · Table 5 · thinking mode unless noted · Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.

Knowledge

Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 Technical Report · Table 12 · thinking · 280 vision tokens
Gemma 4 Technical Report · Table 6 · thinking · 1120 vision tokens
Source: Clinical Benchmarks data docs · Reported by the model's developer. The source relaying it did not verify it. claim · Apr 2026
Setting not stated
MMLU-Pro2 results
Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 Technical Report · Table 5 · thinking mode unless noted
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated

Safety

PhareSource: Phare

Math

AIME 2026Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
MathVision3 results
Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 Technical Report · Table 12 · thinking · 280 vision tokens
Gemma 4 Technical Report · Table 6 · thinking · 1120 vision tokens
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Setting not stated

Long context

RULER 32KGemma 4 Technical Report · Table 9 · without thinking · accuracy · 32K · Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
RULER 128KGemma 4 Technical Report · Table 9 · without thinking · accuracy · 128K · Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.

Factuality

OCR

Other

AIME26Gemma 4 Technical Report · Table 5 · thinking mode unless noted · no tools · Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
ArmBench-LLM Armenian22 related benchmarks · 22 results
BBEHGemma 4 Technical Report · Table 5 · thinking mode unless noted · micro average · Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
GraphWalks · F1Gemma 4 Technical Report · Table 9 · without thinking · F1 · <128K · Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Horangi36 related benchmarks · 72 results
4: ALT_환각방지
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: haerae_bench_v1_rc
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: haerae_bench_v1_wo_rc
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: hrm8k
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: kmmlu
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: kmmlu_pro
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
Show 60 more resultsHide 60 results
4 subset: ko_aime2025
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: ko_arc_agi
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: ko_balt_700_semantic
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: ko_balt_700_syntax
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: ko_hallulens_nonexistent
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: ko_hallulens_wikiqa
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: ko_hle
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: ko_moral
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: ko_truthful_qa
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: kobbq
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: korean_hate_speech
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4 subset: squad_kor_v1
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: ALT_유해성방지
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: ALT_윤리/도덕
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: ALT_제어성
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: ALT_편향성방지
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: FINAL_SCORE
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_구문해석
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_논리적추론
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_수학적추론
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_의미해석
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_일반적지식
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_전문적지식
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_정보검색
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_추상적추론
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_코딩능력
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_표현
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: GLP_함수호출
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: 가치정렬성능(ALT)_AVG
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
4: 범용언어성능(GLP)_AVG
Source: Horangi Korean Leaderboard 4 data docs
Gemma-4-31B-IT-NVFP4: enable-thinking
gemma-4-31B-it: enable-thinking
MTOB4 results
Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 Technical Report · Table 9 · without thinking · full book · ~256K
Gemma 4 Technical Report · Table 9 · without thinking · half book · ~128K
Nejumi 428 related benchmarks · 28 results
Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
Gemma 4 Technical Report · Table 12 · thinking · 280 vision tokens · lower is better
Gemma 4 Technical Report · Table 6 · thinking · 1120 vision tokens · lower is better
Sayf16 related benchmarks · 16 results
Show 10 more resultsHide 10 results
TAU2BenchGemma 4 official model card · aggregate · Reported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.
ValsSource: Vals AI
τ²-benchReported by Google DeepMind · via BenchBoard docs · Reported by the model's developer. The source relaying it did not verify it.

Data credits: Data by LLM Stats · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · SEA-HELM leaderboard, AI Singapore (AI Products team), National University of Singapore. Published aggregate scores; results licence not stated (SEA-HELM code is MIT). · Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · LMArena, CC BY 4.0 (licence) · ArmBench-LLM by Metric-AI. Public numeric primary task results; no separate results licence asserted. · Baromètre de l’écriture IA, Johannes Eva; public numeric ranking data, © 2026 Johannes Eva, no results licence stated · Benchmark Nusantara, ibahasa (M. Khoirul Huda), doi:10.5281/zenodo.21970463. Test results CC BY 4.0, attribution to ibahasa; aggregate results CSV only. (licence) · BullshitBench by Peter Gostev, MIT (licence) · EQ-Bench (Samuel J. Paech) — cite paech2023eqbench · Keelemudelite mõõdupuu: Independent LLM leaderboard for Estonian. Eesti Keele Instituut. https://moodupuu.eki.ee, CC BY 4.0. Latest per-benchmark results selected with partial-run flags retained. (licence) · OpenRouter public benchmark leaderboard and search-lane exports, CC BY 4.0. Fractional source scores converted to percentages; immutable dated snapshots. (licence) · Horangi 4, Weights & Biases Korea; public publisher-selected numeric evaluation results · IdrockBench, IDROCK AI Excellence Center, New Uzbekistan University. Public numeric results; no results licence stated (code MIT and dataset licences not borrowed). · Hugging Face official benchmark leaderboards; licence varies by benchmark · Clinical Benchmarks (clinicalbenchmarks.ai), CC BY 4.0; fractions converted to percentages; source documents retain their own terms (licence) · MineBench, MIT (licence) · Nejumi Leaderboard 4, Weights & Biases Japan; public publisher-selected numeric evaluation results · Outback Yak Research, NepNewsBench v2 research draft. Public numerical aggregates from a candidate preview that the publisher says is not evidence of publication; no release licence assigned and none asserted here. · Phare (Giskard), CC BY 4.0 (licence) · Sayf-eval. Copyright (c) 2026 Qatar Computing Research Institute (QCRI). MIT License. (licence) · Shaberi v2.1, Shisa.AI, Apache-2.0. GPT-5.1-judge scores only; judges are not comparable. (licence) · SnakeBench (Greg Kamradt / ARC Prize Foundation), MIT (licence) · Swallow LLM Leaderboard v2, Swallow LLM team (Institute of Science Tokyo). Contents CC BY 4.0; Japanese post-trained task scores on the source's 0..1 scale. (licence) · UzLiB, Tahrirchi (Abror Shopulatov). Public LEADERBOARD.md numbers; repository MIT covers software and documentation, no separate results licence stated and none asserted. · Vectara Hallucination Leaderboard, Apache-2.0 (licence)

Also known as: aihubmix/gemma-4-31b-it, bedrock_mantle/eu-central-1/google.gemma-4-31b, bedrock_mantle/google.gemma-4-31b, bedrock_mantle/us-gov-west-1/google.gemma-4-31b, cerebras/gemma-4-31b, deepinfra/gemma-4-31b-it, deepinfra/google/gemma-4-31b-it, gemini/gemma-4-31b-it, gemma-4-31b, gemma-4-31b-instruct, gemma-4-31b-it-fp8, gemma-4-31b-it-free, gemma-4-31b-it:free, gemma-4-31b:web, gemma4-31b, gemma4:31b, gmi/google/gemma-4-31b-it, google gemma 4 31b instruct, google-gemma-4-31b-it, google.gemma-4-31b, google/gemma-4-31b, google/gemma-4-31b-it, google/gemma-4-31b-it:free, google/gemma-4-31b-it:thinking, libertai/gemma-4-31b-it, libertai/gemma-4-31b-it-thinking, novita/gemma-4-31b-it, novita/google/gemma-4-31b-it, nvidia_nim/google/gemma-4-31b-it, runware/gemma-4-31b-it, sambanova/gemma-4-31b-it, tee/gemma-4-31b-it, tee/gemma4-31b, tee/gemma4-31b:thinking, tensormesh/google/gemma-4-31b-it, together_ai/google/gemma-4-31b-it, wandb/google/gemma-4-31b-it

Like this one. Tap or hover any dotted value, label or icon to see where it comes from and how it is measured. open details when tapped or hovered. How this data works