Gemma 4 E4B (IT) (Non-reasoning): benchmark results

Provider: Google. Released 2026-04-02. Access: Open.

Unified ELO 1478 ± 1, rank #1674 of 3078 rated models, from 28 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
FrameBench - Japanese84.6Accuracy (%; mean of five prompt templates)52
MERA v2 - GorillaHard20.9Score (%)44.7
MERA v2 - Characters42.7Score (%)36.8
MERA v2 - NewReasoning37.7Score (%)36.8
MERA v2 - Reasoning36.7Score (%)36.8
FrameBench - Frame Identification - Japanese68.5Accuracy (%; Japanese FrameNet candidate frames)36
MERA v2 - Agentic6.7Score (%)35.5
MERA v20.17Total score34.2
MERA v2 - Enantiosemy28.3Score (%)34.2
MERA v2 - IFHardBench6.6Score (%)34.2
MERA v2 - SOB-Hard2.1Score (%)34.2
FrameBench - English82.5Accuracy (%; mean of five prompt templates)31.6

Interactive version: theaggregate.ai/model?slug=gemma-4-e4b-it-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-19.