Gemma 4 12B
Open weightsJun 3, 2026
- Context
- 256K tokens
- Max output
- —
- Input / 1M
- —
- Output / 1M
- —
0 mapped APIs
Model Family
Gemma 4
Gemma 4 is Google DeepMind's current open-weight multimodal model family, spanning edge-focused models, dense local models and a mixture-of-experts variant.
Official documentationVerified Oct 4, 2026
Models
2
Provider
Google
API access
—
Context range
256K tokens
Input price range
—
Output price range
—
Overview
The family supports reasoning, coding, function calling and multimodal understanding. iilib initially represents Gemma 4 with the 12B Unified dense model and the 26B A4B MoE model rather than indexing every size.
Models in this family
Current indexed models and their documented specifications.
Gemma 4 26B A4B
Open weightsApr 2, 2026
- Context
- 256K tokens
- Max output
- —
- Input / 1M
- —
- Output / 1M
- —
0 mapped APIs
| Model | Released | Context | Max output | Input | Output | API |
|---|---|---|---|---|---|---|
| Gemma 4 12B Open weights | Jun 3, 2026 | 256K tokens | — | — | — | — |
| Gemma 4 26B A4B Open weights | Apr 2, 2026 | 256K tokens | — | — | — | — |