BREAKING
Google Adds Two New Gemini Flash Models
3.6 Flash vs 3.5 Flash Lite
Gemini 3.6 Flash
Capable
●
Frontier agentic tasks
●
Sustained coding work
3.5 Flash Lite
Efficient
●
High-throughput execution
●
Subagent workflows
A Quiet Rollout on July 21, 2026
1
Models surface in AI Studio
↓
2
Vertex AI API access
↓
3
Developers test immediately
0
$
per M input tokens
0
$
per M output tokens
0
%
prior Lite speed gain
Specs Still Undisclosed
Flash Line Iterates as Pro Lags
AI NEWS BLITZ
Google has quietly launched two new Gemini Flash models with no formal announcement.