Benchmark menu

Model detail

Gemma 4 31B

A static, shareable model page that remains complete without JavaScript.

Model detail

Gemma 4 31B

Provider: Google · Public slug: google-gemma-4-31b-it

Compare in Charts View game entrants Leaderboard

Rank: 27
Rank movement: Baseline
Score: 24.6
Last measured: 2026-07-05
Coverage: XHigh not supported

Reasoning scores

XHigh: —
Medium: 28.3
None: 20.8

Generation health

Medium0%57%43%n 7None72%14%14%n 7 XHigh not supportedHigh repair rateHigh failure rate

Cost summary

Mean standard list-price equivalent: $0.0045
Median unique-cell cost: $0.0042
Unique-cell range: $0.0014–$0.0092
Coverage: 14 unique cells / 14 observed runs
Observed-run total: $0.06
Evidence basis: Provider-reported
Mean output: 5.5k tok
Output range: 1.3k–13k tok
Token samples: 14

Cost means equal-weight unique model/game/reasoning cells. Provider-account billed spend is reported separately on Charts.

Run setup

Overall Models: combines public reasoning variants
Standard GameBench code-generation prompt