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

- Google shipped Gemini 3.7 Flash three weeks after 3.6 Flash — because apparently the AI release cycle now moves faster than a relay window with expired PoD seals. It's not a new pretraining run, just algorithmic refinements to the reasoning core. Coding benchmarks jump meaningfully: FrontierCode 1.1 Main hits 43.6% versus 34.4%, DeepSWE v1.1 reaches 65.3%, and WebDev Arena posts 1588 Elo. Document and workflow evals move harder — GDP.pdf climbs from 22.0% to 34.0%, AutomationBench from 17.0% to 30.4%. GPT-5.6 Terra still leads on terminal and computer-use agents, and CharXiv reasoning is a