ApertureAug 2026
FieldCompareLabsGuideMethod
  • Field
  • Compare
  • Labs
  • Guide
G

Google DeepMind

London / Mountain View

Gemini 3 Flash line is the speed story of 2026. Pro still trails Claude and GPT on agents; multimodal remains a strength.

deepmind.google
01
G

Gemini 3.7 Flash

Google DeepMind · Aug 2026 · Proprietary

56
Index
Intel56
Code76.1
Agents45.1

Fastest serious reasoning model in the field. GPQA leader on several boards. Agentic depth still behind Claude and Grok.

Price /1M
$0.75 / $3.75
Context
1M
Speed
330 t/s
Status
Current
Low-latency appsMultimodal
Details
02
G

Gemini 3.5 Flash

Google DeepMind · Jun 2026 · Proprietary

52
Index
Intel52
Code68
Agents41

Still widely deployed. Intelligence is in the Flash cluster; price is worse than 3.7.

Price /1M
$1.50 / $9
Context
1M
Speed
191 t/s
Status
Superseded
Existing Gemini 3.5 traffic
Details
03
G

Gemini 3.6 Flash

Google DeepMind · Jul 2026 · Proprietary

51.6
Index
Intel51.6
Code69.2
Agents40.5

Last month’s Flash. 3.7 is faster, cheaper, and higher. Only keep it if an endpoint is pinned.

Price /1M
$1.50 / $7.50
Context
1M
Speed
238 t/s
Status
Superseded
Pinned Gemini endpoints
Details
04
G

Gemini 3.1 Pro

Google DeepMind · May 2026 · Proprietary

47.7
Index
Intel47.7
Code68.8
Agents23

Google’s Pro that actually shipped. Knowledge and multimodal are real; agentic scores are the hole in the card.

Price /1M
$2 / $12
Context
1M
Speed
80 t/s
Status
Current
Multimodal analysisLong documents
Details

Aperture is a field guide, not a vendor. Scores compiled 28 Aug 2026.

MethodologyHow to pick