Sulat.com
AI models
$10 off the fastest DeepSeek V4.1 Flash, Kimi K3 and GLM 5.3 from Synthetic
CrossModel logo

Model details

Gemini 3.7 Flash

Gemini 3.7 Flash is the latest iteration in Google's Flash-tier line, arriving only about three weeks after the prior 3.6 Flash release in a notably accelerated cadence. Google DeepMind positions it as "our most intelligent workhorse model yet for coding and agents," with a tagline framing it as "best for tackling complex agentic tasks at scale." The company publicly characterized the jump from 3.6 to 3.7 as offering "substantial improvements," signaling that even within the lightweight Flash family this release was treated as a meaningful capability step rather than a routine patch.

In practical terms, Gemini 3.7 Flash is aimed at developers who need agentic reasoning and coding assistance without paying top-tier latency costs, sitting in the same Flash family slot that has traditionally traded some raw intelligence for speed and throughput. Google makes it reachable through two clear entry points: a consumer-facing "Try in Gemini" link and a "Build with Gemini" path in Google AI Studio that targets the gemini-3.7-flash identifier, so teams can prototype agentic workflows quickly before deeper integration. The combination of an agent-focused design intent, an unusually fast release cycle, and dual consumer plus developer availability makes it a sensible fit for high-volume coding assistants, tool-using agents, and other production systems where Flash-tier latency is more valuable than flagship-tier reasoning depth.

CrossModelgemini/gemini-3.7-flashgemini-flash

Quick Info

Powered by
Provider
CrossModel
Model key
gemini/gemini-3.7-flash
Release date
Aug 13, 2026
Last updated
Aug 13, 2026
Knowledge cutoff
2026-03
Input modalities
Output modalities
Capabilities

Cost

Input token cost
$0.75
Output token cost
$3.75

Limits

Output tokens
65,536 tokens
Context window
1,048,576 tokens

Transparent token rates

Compare Gemini 3.7 Flash pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about Gemini 3.7 Flash

CrossModel

Coverage

NYU Shanghai RITS analyzes Gemini 3.7 Flash as a three-week point release built on the same pretrained foundation as Gemini 3.6 Flash, with Google's model card explicitly describing it as "algorithmic improvements to its core reasoning foundation" rather than a new pretraining run. The piece retains 3.6 Flash's specs, The analysis also documents gains outside pure coding, including GDP.pdf rising from 22% to 34%, Harvey LAB-AA complex legal workflows improving from 85.1% to 90.7%, and MRCR v2 128k long-context retrieval moving from 91.8% to 97.0%, while noting that some non-coding evaluation deltas were flat or moved backward. The p

CrossModel

Coverage

EvoLink confirms Gemini 3.7 Flash reached general availability on August 13, 2026 under the stable model ID "gemini-3.7-flash" with no preview suffix. Specifications verified against Google's documentation include a 1,048,576-token context window, 65,536-token maximum output, multimodal input support (text, image, vide Official introductory pricing is $0.75 input / $3.75 output per 1M tokens through December 31, 2026, stepping up to $1.50 / $7.50 from January 1, 2027, with thinking tokens billed at the output rate. EvoLink notes that the upgrade case is capability and token efficiency rather than a lower rate card, since 3.6 Flash ca

CrossModel

Coverage

Ars Technica reports that Google announced Gemini 3.7 Flash on August 13, 2026, rolling it out to replace Gemini 3.6 Flash just three weeks after that predecessor's release. Google's Senior Director Tulsee Doshi framed the new model as a "workhorse" focused on coding and agentic performance, citing gains on FrontierCod Ars Technica questions whether the benchmark deltas alone justify a new release three weeks after 3.6 Flash and situates the launch within a broader 2026 slowdown in Google's flagship cadence, noting that the long-promised Gemini 3.5 Pro never materialized on its announced June timeline. The piece characterizes the new

CrossModel

Coverage

Digital Applied's independent analysis published August 13, 2026 highlights a critical pricing nuance largely missed by other launch coverage: while Google marketed 3.7 Flash at half the workhorse tier's cost, the company simultaneously cut Gemini 3.6 Flash to the identical $0.75/$3.75 per million token rate, so the ha The piece grounds its capability assessment in independent measurement, citing a +4 gain on the Artificial Analysis Intelligence Index (56 vs 52 for 3.6 Flash), and notes Google's own model card shows 3.7 Flash winning 7 of 13 benchmarks but losing 4 to GPT-5.6 Terra. The analysis frames the launch as occurring 23 days

CrossModel

Coverage

Google announced Gemini 3.7 Flash on August 13, 2026, describing it as "our most intelligent workhorse model yet for coding and agents," arriving just three weeks after Gemini 3.6 Flash as a direct response to developer feedback and algorithmic innovations. The official Google blog post, authored by Tulsee Doshi (Senio The post reports notable benchmark gains over 3.6 Flash, including FrontierCode 1.1 Main rising to 43.6% (from 34.4%), DeepSWE v1.1 reaching 65.3% (from 49.0%), GDP.pdf climbing to 34.0% (from 22.0%) for complex document processing, AutomationBench reaching 30.4% (from 17.0%) for real-world business workflows, and WebD

Videos about Gemini 3.7 Flash

More models around Gemini 3.7 Flash