AILet’s talk
Back to AI updates
GoogleModel releases

Introducing Gemini 3.7 Flash

If your team uses AI for coding, document review, or building small internal tools, Gemini 3.7 Flash offers noticeably better first-pass code accuracy and document handling at half the previous price. Worth a look if you already use Gemini for dev or finance/legal document work, or if you are choosing a lower-cost model tier for automations.

Read the original source 3 min read

What changed

Source summary

Google released Gemini 3.7 Flash, a faster workhorse model for coding and agentic tasks, with notable gains on coding benchmarks (FrontierCode 1.1 Main and DeepSWE v1.1) and improved reasoning for finance, law, and biosciences document work. Introductory pricing is $0.75 per 1M input tokens and $3.75 per 1M output tokens through the end of 2026 — half the cost of Gemini 3.6 Flash. It is rolling out in the Gemini app (Spark), Google AI Studio, Android Studio, and enterprise platforms.

Why it matters to your work

Practical interpretation

If your team uses AI for coding, document review, or building small internal tools, Gemini 3.7 Flash offers noticeably better first-pass code accuracy and document handling at half the previous price. Worth a look if you already use Gemini for dev or finance/legal document work, or if you are choosing a lower-cost model tier for automations.

When this could help

Worth switching to now if you use the Gemini API for coding or agentic tasks — introductory pricing ($0.75/$3.75 per 1M input/output tokens) runs through Dec 31 2026, doubling Jan 1 2027.

Gemini 3.7 Flash is Google's most capable Flash-tier model, optimised for coding and agentic workloads. Versus 3.6 Flash: FrontierCode 43.6% (vs 34.4%), DeepSWE 65.3% (vs 49.0%), WebDev Arena Elo 1588 (vs 1538), AutomationBench 30.4% (vs 17.0%). Better at debugging, multi-step planning, producing production-ready code first try, and matching reference designs from screenshots.

How to get started

  1. Access via Google AI Studio (developers), the Enterprise Agent Platform (enterprises), or Gemini Spark (Pro/Ultra subscribers). Select Gemini 3.7 Flash as your model.

    Provider guidance ↗

Build it into your routine

SCALA suggested workflow

  1. Coding agent upgrade

    If you run coding or agentic workloads on Gemini 3.6 Flash, switch to 3.7 Flash for the benchmark improvements — especially on DeepSWE (+16.3pp) and AutomationBench (+13.4pp) — while introductory pricing lasts.

What to check

  • All benchmark comparisons are against 3.6 Flash only — position vs Pro or other tiers is not stated. Introductory pricing doubles Jan 1 2027. Source ↗
  • Spark consumer access requires a paid subscription and is restricted in the EEA, Nigeria, Switzerland, and the UK. Source ↗

Sources & provider guides

Official documentation checked 12 Oct 2026. SCALA prompts and workflows are suggestions, not claims of hands-on testing.

Source & availability

This is a summary of a published provider announcement, with practical context from SCALA’s feed.

Plan, region, workspace and rollout conditions may affect access in Australia. Check the original source for current availability.

blog.google