Google Releases Gemini 3.7 Flash, Cuts Price in Half Versus 3.6 Flash
Google has released Gemini 3.7 Flash, just three weeks after Gemini 3.6 Flash, claiming substantial gains in coding, web development, and document reasoning. The model launches at an introductory price of $0.75 per 1M input tokens and $3.75 per 1M output tokens — half the cost of its predecessor.
Google DeepMind released Gemini 3.7 Flash on August 13, 2026, positioning it as its "most intelligent workhorse model yet for coding and agents." The release comes just three weeks after Gemini 3.6 Flash, continuing Google's rapid iteration cadence on the Flash line.
The model launches at an introductory price of $0.75 per 1M input tokens and $3.75 per 1M output tokens — half the per-token cost of 3.6 Flash, according to Google. That pricing is confirmed as an introductory rate available through the end of the year, with future pricing unspecified.
Benchmark claims
Google reports the following gains for 3.7 Flash over 3.6 Flash, based on internal and third-party benchmarks:
- FrontierCode 1.1 Main: 43.6% vs 34.4%
- DeepSWE v1.1: 65.3% vs 49.0%
- WebDev Arena (Arena.ai) Elo: 1588 vs 1538
- GDP.pdf (complex document processing): 34.0% vs 22.0%
- AutomationBench (real-world business workflows): 30.4% vs 17.0%
Google claims the model shows stronger first-pass code accuracy, better production-ready code generation, and improved design adherence when generating UIs from screenshots or design systems. In knowledge-dense domains like finance, law, and biosciences, the company says 3.7 Flash delivers materially better reasoning and document-processing accuracy than its predecessor. None of these benchmark figures have been independently verified by third parties as of publication.
Developer experience and rollout
According to Google, 3.7 Flash better handles roadblocks, clarifies ambiguous intent, and follows instructions with higher fidelity than 3.6 Flash, while dedicating more computation to multi-step planning and tool calls. The company frames this as reducing manual oversight and retries in agentic coding workflows.
The model is available immediately through the Gemini API in Google AI Studio, Android Studio, and Google Antigravity for developers; through the Gemini Enterprise Agent Platform and Gemini Enterprise app for businesses; and through Gemini Spark — Google's 24/7 personal AI agent — for Google AI Pro and Ultra subscribers in more than 160 countries. Spark, launched at Google I/O, is being upgraded to run on 3.7 Flash starting today, with Google citing improved tool use across Google Workspace apps for tasks like drafting emails and consolidating files.
Google also states that 3.7 Flash ships with updated Frontier Safety safeguards targeting misuse in CBRN (chemical, biological, radiological, nuclear) domains and cyber offense, consistent with its existing bioresilience and cyber safety programs. Full technical specifications, including exact context window size and parameter count, were not disclosed in the announcement; the company points to a separate model card for additional detail.
What this means
A three-week gap between Flash releases signals Google is now iterating on its lower-cost model tier at a pace closer to weekly software updates than traditional model-release cycles. Halving the price while claiming double-digit benchmark improvements on coding and business-automation tasks is an aggressive move against competitors' cheaper tiers, particularly for developers running high-volume agentic workloads where token costs compound quickly. The benchmark gains, if they hold up under independent testing, would matter most for teams already building coding agents and document-processing pipelines on Gemini Flash — but until third-party evaluation confirms the FrontierCode, DeepSWE, and AutomationBench numbers, these remain Google's own claims. The tight release cadence also raises questions about how much genuine architectural change versus fine-tuning is driving each successive Flash version.
Related Articles
Google Launches WeatherNext 3, Claims 50% More Accurate Precipitation Forecasts
Google DeepMind and Google Research released WeatherNext 3, a weather AI model trained on real-time geostationary satellite data instead of lagging numerical weather prediction outputs. Google claims up to 50% more accurate day-ahead precipitation forecasts, now rolling out to Search, Maps, and the Gemini app.
Google Launches Gemini 3.8 Flash With Same Pricing as Predecessor, Its Third Flash Model in Six Weeks
Google released Gemini 3.8 Flash, its third Flash model in six weeks, pricing it identically to its predecessor at $0.75 per million input tokens and $3.75 per million output tokens. The launch coincides with a favorable antitrust ruling and public praise from Berkshire Hathaway's Greg Abel, giving Alphabet a stronger narrative after a four-month stock slide.
OpenAI's GPT-6 Astra Reportedly Automates AI Engineering Tasks at Under $6 an Hour, According to Latent Space Testing
A Latent Space report describes GPT-6 Astra, a new OpenAI model the blog says can autonomously handle AI engineering tasks—training models, labeling data, deploying systems—at an estimated cost of under $6 per hour. The claims, including 97.6% on FrontierMath and 99.9% on ARC-AGI-3, come from independent blog testing rather than an official OpenAI announcement.
Google DeepMind Launches WeatherNext 3, Cutting Forecast Grid to 5km Using Live Satellite Data
Google DeepMind and Google Research released WeatherNext 3, a global weather AI model that ingests live geostationary satellite data to produce hourly forecasts at up to 5-kilometer resolution — five times sharper than its predecessor, WeatherNext 2. The model is now integrated into Google Search, Maps, Gemini, Google Maps Platform, and Cloud.
Comments
Loading...