Gemma 4 update delivers Flash Attention 4 for Nvidia H100 users, fixes tool-calling bugs, and expands vision resolution — all under the same model name with no version bump. If your weights predate July 16, 2026, re-pulling from Hugging Face activates the new inference speed gains and agentic reasoning improvements.
Gemma 4 Gets Stealth Update: H100 Speed Gains and Tool-Calling Fixes, No Version Bump
