Google Releases Gemini 3.7 Flash: A Coding and Agent Model at $0.75/1M Input Tokens

Google released Gemini 3.7 Flash three weeks after Gemini 3.6 Flash, describing it as a refinement of the earlier model with algorithmic improvements to core reasoning. The model handles text, images, audio, and video within a 1M-token context window and returns up to 64K output tokens. Key performance gains are in software engineering, document-heavy tasks, and web development. On FrontierCode 1.1 Main, it scores 43.6% versus 34.4% for 3.6 Flash; on DeepSWE v1.1, it reaches 65.3%; on WebDev Arena, it achieves an Elo of 1588. Document and workflow benchmarks improved significantly: GDP.pdf went from 22.0% to 34.0%, and AutomationBench from 17.0% to 30.4%. GPT-5.6 Terra still leads on some benchmarks, and CharXiv Reasoning shows a slight regression. Pricing is introductory: $0.75 per 1M input tokens and $3.75 per 1M output tokens until December 31, 2026, then it doubles. The model is available only via API and enterprise platforms, with no open weights or self-hosting options.