Gemini 3.7 Flash
Google · Released Aug 2026
Gemini 3.7 Flash is Google's strongest Flash workhorse yet for coding and agents — a genuine step up from 3.6 Flash at half the prior price through 2026. It still trails the frontier on the hardest terminal and computer-use tasks. A clear upgrade for cost-sensitive and Google-ecosystem agent builders; not a frontier-code replacement.
Is it right for you?
Good for
- Cost-sensitive coding and agent workloads — $0.75/$3.75 intro pricing is roughly a third the blended cost of Claude Sonnet 5 or GPT-5.6 Terra
- Long-horizon software engineering — DeepSWE v1.1 65.3%, up from 49.0% for 3.6 Flash
- Web development — Code Arena Elo 1588, top of Google's comparison table, with better design-system adherence in fewer prompts
- Google-ecosystem agent builders — powers Gemini Spark; available via Gemini API, AI Studio, Android Studio, and Antigravity
Not good for
- Hardest terminal/agentic coding where frontier models lead — GPT-5.6 Terra is ahead on DeepSWE (69.6%), Terminal-Bench 2.1 (87.4%), and OSWorld-2.0
- Teams needing price stability beyond 2026 — the introductory rate expires Dec 31, 2026, then doubles to $1.50/$7.50
How it performs by task
Software engineering (DeepSWE v1.1)
65.3%, a big gain over 3.6 Flash's 49.0% but still behind GPT-5.6 Terra's 69.6%
Production code quality (FrontierCode 1.1 Main)
43.6% vs 34.4% for 3.6 Flash; leads GPT-5.6 Terra on Google's table
Terminal/CLI agent work (Terminal-Bench 2.1)
85.8%, close behind GPT-5.6 Terra's 87.4%
Web development (Code Arena)
1588 Elo — top score in Google's comparison table
Knowledge work / long context
1M context, 97.0% on the 128K long-context needle test; GDPval-AA 1525 Elo trails Sonnet 5's 1598
Pricing
Input
$0.75 / 1M
Output
$3.75 / 1M
Context
1M context, 64K output
Benchmarks
No verdict changes yet
The clock starts day one. Changes land here as our verdict evolves.
Sources
Verification log
No verification checks yet
We haven't logged a verification check for this entry. Once a check runs, its history shows here.