Google is keeping up a super-fast pace with its AI updates. Just three weeks after releasing Gemini 3.6 Flash, the company is already rolling out Gemini 3.7 Flash. Built as a quick, reliable workhorse for developers, this new version lands as a direct response to user feedback. It brings noticeable speed and accuracy boosts for programming, handling multi-step tasks, and running automated agents.
To get more people using it, Google is cutting introductory pricing in half through the end of the year. Gemini 3.7 Flash now costs $0.75 per million input tokens and $3.75 per million output tokens. That gives developers a way cheaper option compared to competing models like OpenAI’s GPT 5.6 Luna while still delivering solid performance.
Big gains in coding, complex documents, and safety
Despite coming out so quickly after the last version, the benchmarks show some pretty impressive jumps. For coding, the model made a huge leap on the DeepSWE v1.1 test, jumping from 49.0% to 65.3% for finding bugs and resolving code issues. It also bumped its FrontierCode 1.1 Main score from 34.4% to 43.6%, while its WebDev Arena rating rose to 1,588, meaning it can generate working web layouts and full app designs with fewer prompts and better UI design adherence.
Knowledge work gets a nice upgrade too. On the GDP.pdf test, which checks how well an AI reads through complex documents like legal or financial files, Gemini 3.7 Flash scored 34.0% compared to 3.6 Flash’s 22.0%. It also hit 30.4% on AutomationBench for handling real-world business workflows. Google says the model is much better at adapting when it hits a roadblock, asking for clarification when needed, and following instructions without requiring constant retries. On top of that, Google added updated safety guardrails against CBRN (chemical, biological, radiological, and nuclear) misuse and cyber threats (via 9to5Google).
Where to try it and the status of Gemini 3.5 Pro
If you want to test it out, Gemini 3.7 Flash is available today across the Gemini API, Google AI Studio, Android Studio, Google Antigravity, and the Gemini Enterprise platform. Individual users will find it powering the Gemini Spark agent inside the main Gemini app. It currently requires an AI Pro or Ultra subscription, though. However, the updated model is available in 160 countries. Free users and AI Plus subscribers stay on Gemini 3.6 Flash for now.
The launch comes at an interesting time for Google. The company still hasn’t released its flagship Gemini 3.5 Pro model, which was originally teased earlier this summer. Plus, there’s a major leadership shuffle at Google DeepMind that saw CEO Demis Hassabis step aside for Koray Kavukcuoglu.
Co-founder Sergey Brin has urged staff to focus heavily on Gemini’s core development. However, rumor has it Google might skip 3.5 Pro altogether and jump straight to training Gemini 4.
The post Gemini 3.7 Flash Launches with Major Performance Gains and Half-Off Pricing appeared first on Android Headlines.
​Â
