gemma-4-31B-it-FP8-dynamic: benchmark results

Provider: Google. Released 2026-04-04. Access: Open.

Aggregate Rating 1741 ± 16, rank #131 of 1635 rated models, from 213 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
ObligBench - Control - Intention Planning - BVA At Inferred Bound0.97Conditional native adequacy fraction (0-1)100
ObligBench - Control - Intention Planning - BVA Empty0.8Conditional native adequacy fraction (0-1)100
ObligBench - Control - Intention Planning - BVA False Value1Conditional native adequacy fraction (0-1)100
ObligBench - Control - Intention Planning - Oracle Return Behavior1Conditional native adequacy fraction (0-1)100
ObligBench - Control - Intention Planning - Oracle Return Boolean1Conditional native adequacy fraction (0-1)100
ObligBench - Control - Intention Planning Execution100Generated test execution pass rate (%)100
ObligBench - Control - Mutation Feedback - BVA Above Inferred Bound0.97Conditional native adequacy fraction (0-1)100
ObligBench - Control - Mutation Feedback - BVA Empty0.78Conditional native adequacy fraction (0-1)100
ObligBench - Control - Mutation Feedback - CFA Else If0.47Conditional native adequacy fraction (0-1)100
ObligBench - Control - Mutation Feedback Execution99.86Generated test execution pass rate (%)100
ObligBench - Control - Panta - BVA At Inferred Bound0.9Conditional native adequacy fraction (0-1)100
ObligBench - Control - Panta - BVA Empty0.65Conditional native adequacy fraction (0-1)100

Interactive version: theaggregate.ai/model?slug=gemma-4-31b-it-fp8-dynamic · How It Works · Data refreshed daily, snapshot 2026-11-05.