$0.05 per 1M input tokens and $0.10 per 1M output tokens (cheapest host: Novita).
There is no free API; the cheapest paid price is $0.05 per 1M input tokens. The weights are open, so you can run it yourself.
#318 of 360 in Chat & Knowledge, #134 of 138 in Agents & Tool Use, #111 of 197 in Writing & Language, #126 of 132 in Vision & Documents.
On MMMU · validation, Gemma 3 12B scores 59.6 vs 60.3 — about the same.
Through the API of AWS Bedrock, DeepInfra, NextBit, Novita.
131K tokens, with up to 16K output tokens.
Released 2025-03-01. It is the current version in its line.
Yes, the weights are open (gemma licence).
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
It accepts text and images as input.
Gemma 3 12B is an open-weights language model from Google, released Mar 2025.
GPT-5.1 (better), MiMo VL 7B RL (open weights).