Gemma 4 12B IT
Latest GemmaCompact Gemma 4 instruction model for open, self-hosted chat and reasoning
- Strongest at
- Writing Better than 30% of the past year's models · 2 of 4 benchmarks
- Context window
- 262.1K tokens in · 32.8K out
- Since the last one
- 68 days after Gemma 4 31B IT
Capability profile
Where it lands among 91 models released in the past year. 50 is the average model.
This modelGemma 4 31B ITEvery model from the past year
Coding
No results yet
Writing
49Better than 30%
Reading & documents
No results yet
Reasoning
No results yet
Math
No results yet
Science
No results yet
Visual reasoning
No results yet
Computer use
No results yet
Where these scores come from
Each capability combines the published benchmarks listed here, rescaled so 50 is the reference average and 15 points is one standard deviation. A hollow marker rests on a single benchmark.
Since Gemma 4 31B IT
6 releases in the Gemma line.
Gemma 2 2B ITJul 31, 2024224dGemma 3 12B ITMar 12, 20250dGemma 3 27B ITMar 12, 20250dGemma 3 4B ITMar 12, 2025386dGemma 4 31B ITApr 2, 202668dGemma 4 12B ITJun 9, 2026
Released68 days later
Context262.1Ksame
Writing49−4.3
Google's past year
37 release days in the past year, about one every 10 days. The next is expected around Nov 3, 2026. Forecasts ↗
NovDecJanFebMarAprMayJunJulAugSepOctNov
Specifications
RELEASE DATEJun 9, 2026
WEIGHTSOpen weights
CONTEXT WINDOW262.1K
MAX OUTPUT32.8K
REASONINGNot listed
TOOL CALLINGNot listed
INPUTStext, image/jpeg, image/png
OUTPUTStext
Specifications and prices from Models.dev; availability and limits may differ by provider. Benchmarks from Epoch AI (CC BY 4.0), LiveBench and the other sources named above.