Skip to content
aiqbench

Google · model passport

Gemma 4 31b

gemma-4-31b
  • Open weights
  • Apache 2.0 licence
  • 32.7B params
  • Released 2026-03-12(HF repo created)
9/27axes covered
Overall · LMArena
1,452#67 of 396
Best percentile
92.0IFBench (Artificial Analysis)
Cohort standing
64.9median 57.2

What the data says

Benchmarked on 9 of 27 tracked axes. Every figure below is the model's own published result on that board.

Strongest axes

  • 92 pct — IFBench (Artificial Analysis) (75.6%)
  • 87.4 pct — Instruction following (1,454)

Uneven: 65.7 percentile points between its best and worst axis — 92 pct on IFBench (Artificial Analysis) versus 26.3 pct on Document Arena.

  • IFBench (Artificial Analysis)92 pct75.6%Verify →
  • Instruction following87.4 pct1,454Verify →
  • Hard prompts84.3 pct1,476Verify →
  • Coding83.9 pct1,502Verify →
  • Overall83.3 pct1,452Verify →
  • Vision Arena76 pct1,262Verify →
  • Factuality Arena65.1 pct1,449Verify →
  • WebDev Arena30.2 pct1,368Verify →
  • Document Arena26.3 pct1,445Verify →
  • Search ArenaNot benchmarked
  • Agent ArenaNot benchmarked
  • GPQA DiamondNot benchmarked
  • AIME (math)Not benchmarked
  • FrontierMathNot benchmarked
  • SimpleQA VerifiedNot benchmarked
  • ARC-AGI-2Not benchmarked
  • Humanity's Last ExamNot benchmarked
  • SimpleBenchNot benchmarked
  • LiveBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • MMLU-Pro (Artificial Analysis)Not benchmarked
  • LiveCodeBench (Artificial Analysis)Not benchmarked
  • HHEM Hallucination (Vectara)Not benchmarked
  • SWE-bench VerifiedNot benchmarked
  • Terminal-BenchNot benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

How to run it

Run it yourself

Repositorygoogle/gemma-4-31B
Parameters32.7B
ArchitectureGemma4ForConditionalGeneration
LicenceApache 2.0

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "google/gemma-4-31B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

Ollama

We don't sync Ollama tags, so we won't print a command we can't verify — a wrong one would pull the wrong model. Search the Ollama library →

Unsloth fine-tuning recipes are model-specific and we don't derive them here. Unsloth docs →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

  • gemma-4-31b stands at 64.9 (raw mean 69.8) — measured on 9 of 27 axes.

    standingtemplateVerify →

  • gemma-4-31b stands at 64.9 versus a cohort median of 57.2 across 101 models released within 180 days — above the peer median.

    cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →