BREAKING
Google Ships Gemma 4 Update
0%
max prefill gain
0%
faster TTFT
0x
tokens per sec via MTP
Practical Fixes for Production
1Tool calling accuracy
2Cleaner chat template
3Fix laziness
Vision Detail vs Efficiency
Default280 tokens
Efficient processing
Lower cost
Max fidelity1,120 tokens
Sharper OCR
Images up to 2.51MP
Reliability Is the New Battleground
Updated Weights Live on Hugging Face
AI NEWS BLITZ
Google is rolling out a major update to its open-weight Gemma 4 model family.