Google DeepMind·· 2026-04-03
Gemma 4: Byte for byte, the most capable open models
Gemma 4: Byte for byte, the most capable open models
AI summary
Google DeepMind released the open-source Gemma 4 family in four sizes—E2B, E4B, 26B MoE and 31B Dense—under Apache 2.0.
Selection record
Threshold 60Official, first-handFirst 82Second 82
AdmittedSum of both 164 ≥ twice the threshold 120
- Source tier
- Official, first-hand; this tier's threshold is 60
- Pre-filter
- passed：发布Gemma 4开源模型，含推理与Agent能力
- Why it was chosen
- The four sizes and Apache 2.0 licence clarify the availability of open-source models for local and on-device deployment.
A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.
Source: Google DeepMind · deepmind.google