BREAKING
Google Adds 3 New Gemini Flash Models
Three Flash Tiers, Three Jobs
3.6 Flash
GA
●
Workhorse model
●
Coding & multimodal
3.5 Flash-Lite
GA
●
High-throughput
●
Low cost everyday
3.6 Flash Beats 3.5 on Benchmarks
DeepSWE 3.5
37
DeepSWE 3.6
49
MLE 3.5
49.7
MLE 3.6
63.9
0
%
more token efficient
0
$
3.6 input per M tokens
0
/s
Flash-Lite tokens
Wide Rollout, One Limited Model
1
3.6 Flash GA
↓
2
Flash-Lite GA
↓
3
Cyber restricted
Google Bets on the Flash Tier
AI NEWS BLITZ
Google DeepMind is expanding its Gemini family with three new Flash models built for efficiency.