Google DeepMindGA
Gemini 3.7 Flash
Google’s current Flash tier — a coding and agent refresh at an introductory price with a known expiry.
Context
1.0M
Max output
64K
Input /1M
$0.38
Output /1M
$1.88
Live pricing via OpenRouter
Best for
- High-volume agentic coding where cost compounds
- Web development and agent workflows (Google’s stated focus)
- 1M-context work at sub-dollar input pricing
Watch out
The $0.75/$3.75 rate is introductory through 2026-12-31 and doubles to $1.50/$7.50 on 2027-01-01 — do not build a cost model on it without noting the step. Gains over 3.6 Flash are vendor-reported and unreproduced independently.
For creators. The cheap agent runtime for bulk drafting and content-generation loops, with a calendar reminder for the price change.
Benchmarks
| deepswe v1 1 | 65.3 |
| automationbench | 30.4 |
| webdev arena elo | 1588 |
Capabilities
- 1,048,576-token context, 64K max output
- Positioned by Google for software engineering, web development and agent workflows
- DeepSWE v1.1 65.3%, up from 49.0% on Gemini 3.6 Flash (vendor-reported)
- Introductory $0.75/$3.75 pricing through 2026-12-31