Google · Model release
Gemini 3.7 Flash: the August 2026 coding and agent release
Gemini 3.7 Flash arrived three weeks after 3.6 Flash as Google’s Flash workhorse for coding and agents, with introductory API rates through the end of 2026. Gemini 3.8 Flash (2026-09-02) is newer; compare both on the target workload instead of inferring fit from release order.
What changed
- Google shipped a new Flash id, `gemini-3.7-flash`, as GA on 2026-08-13.
- Paid introductory rates are $0.75 / $3.75 per million tokens through 2026-12-31, then $1.50 / $7.50.
- In the dated public snapshots reviewed on 2026-10-01, DeepSWE 1.1 lists best-effort at high effort (65% at $2.18/task) and the rebased AA Intelligence Index v4.3.2 lists 39.1 at high. CursorBench High 61.6% was on version 3.2; CursorBench 4.0 does not list 3.7 Flash. These are source-, effort-, and harness-specific rows, not a current universal ranking.
Best fit
- Coding agents
- Multimodal agent loops
- High-volume Google API work
What to be careful about
- Introductory pricing expires on 2026-12-31.
- DeepSWE Best uses the high-effort row (65.3%), not the slightly higher medium score.
- Max output stays 65K tokens.
Dataset snapshot: Gemini 3.7 Flash is also represented in the checked comparison dataset. Prices and benchmark figures carry their own field dates.