Google launched Gemini 3.7 Flash on August 13, 2026, scoring 43.6% on the FrontierCode 1.1 Main benchmark — above Claude Sonnet 5’s 42.7% — and priced it at $0.75 per million input tokens through December 31, 2026, half the rate of its predecessor Gemini 3.6 Flash. The model is live in Gemini Spark and available via the Google AI API.
The launch makes Gemini 3.7 Flash the most cost-performance-competitive model Google has shipped for coding and AI agent workloads in the Flash tier, arriving three weeks after Gemini 3.6 Flash in a cadence that has now produced three Flash-tier releases in roughly 16 weeks.
Gemini 3.7 Flash Benchmark Scores Across All Five Tests
Gemini 3.7 Flash improved across every benchmark tracked from Gemini 3.6 Flash. Google attributes the gains to algorithmic changes in the model’s reasoning foundation rather than an increase in parameter count, which was not disclosed.
- FrontierCode 1.1 Main: 43.6% — above Claude Sonnet 5 (42.7%), which had held the top score for production code quality in Q2–Q3 2026.
- DeepSWE v1.1: 65.3% — up from 49.0% on Gemini 3.6 Flash, a 16.3-percentage-point gain.
- AutomationBench: 30.4% — up from 17.0%, a 13.4-percentage-point improvement.
- GDP.PDF document comprehension: 34.0% — up from 22.0% on the previous generation.
- Arena.ai WebDev Elo: 1588 — placing it above competing models in web development task evaluation.
Google stated the model outperforms comparable Anthropic and OpenAI models across 9 benchmarks total, per the official announcement and coverage by SiliconAngle and OfficeChai. The FrontierCode and DeepSWE scores are the most directly relevant for businesses running coding agents or automated software workflows.
What Gemini 3.7 Flash Costs — Intro Rate vs Standard Rate
Gemini 3.7 Flash costs $0.75 per million input tokens and $3.75 per million output tokens through December 31, 2026. From January 1, 2027, the rate reverts to $1.50 per million input tokens and $7.50 per million output tokens — the same standard rate that Gemini 3.6 Flash carried at launch.
The introductory rate is 75% lower per input token than Claude Sonnet 5’s current promotional price of $3 per million input tokens, which applies through August 31, 2026. After August 31, Sonnet 5 continues at $3/$15 as its standard rate. The price gap between Gemini 3.7 Flash (intro) and Sonnet 5 (standard) is 4× on input tokens.
The model supports a 1-million-token input context window and generates up to 64,000 output tokens per request.
Why the Timing Matters for Businesses Evaluating AI Agents
Flash-tier models power the majority of production AI agent deployments because they process the high request volumes that agent loops require at a cost sustainable for automation at scale. Gemini 3.7 Flash’s FrontierCode 1.1 Main win over Claude Sonnet 5, combined with a 75% price advantage under the intro rate, positions it as the primary alternative for teams reviewing their AI coding and orchestration stacks before the August 31 Sonnet 5 promotional deadline. For an overview of where these models fit, see our guide to AI agents for business tasks.
Google’s release timing appears deliberate: shipping a benchmark-leading model at half-price in the same week that a competitor’s discount expires is a direct bid for API contract switches, specifically in the segment of teams running cost-sensitive agent infrastructure at volume. Businesses that lock in API usage under the intro rate retain the $0.75/M input price through December 2026; the January 2027 rate doubles. Evaluating model fit for agent workflows is covered in our guide to the best AI tools for business in 2026.
For Context: Gemini and AI Agent Coverage on WithO2
- Gemini’s leadership transition — what the DeepMind CEO change signals for Google’s AI model roadmap.
- AI agents in practice — 15 real-world business use cases where Flash-tier models like Gemini 3.7 run the core logic.
- Agent orchestration frameworks — YC’s open-source harness and how it integrates with cost-optimised foundation models.
Google is running a pricing war dressed as a benchmark release. Shipping Gemini 3.7 Flash at exactly half Gemini 3.6 Flash’s price in the same week Anthropic’s Sonnet 5 promotional rate expires is not coincidence — it is a deliberate window to capture API contract switches from cost-sensitive teams. The FrontierCode margin over Sonnet 5 is narrow (43.6% vs 42.7%), so the economics, not the scores, are the actual story. Businesses should lock in usage under the $0.75/M input rate before it doubles on January 1, 2027.
