Google releases Gemini 3.7 Flash, priced at half of 3.6 Flash
On Aug 13 Google launched Gemini 3.7 Flash targeting coding and agents, with major coding benchmark gains and year-end API pricing of $0.75/$3.75 per million tokens.
On August 13, Google DeepMind released Gemini 3.7 Flash, a workhorse model targeting coding and agent workloads, launched just three weeks after Gemini 3.6 Flash. The model supports a 1,048,576-token input context, 65,536-token output limit, and configurable thinking modes at low, medium, and high settings (the minimal setting is explicitly unsupported and returns an error).
Benchmarks show significant gains over 3.6 Flash: DeepSWE v1.1 coding improved from 49.0% to 65.3%, FrontierCode 1.1 Main from 34.4% to 43.6%, and WebDev Arena from 1538 to 1588. On knowledge work, the GDP.pdf benchmark rose from 22.0% to 34.0%, and AutomationBench from 17.0% to 30.4%.
API pricing is $0.75 per million input tokens and $3.75 per million output tokens—half the launch price of 3.6 Flash—effective through year-end. However, OpenAI GPT-5.6 Luna is priced even lower at $0.20 input and $1.20 output per million tokens, underscoring the intensifying AI price war. Gemini 3.7 Flash is available via the Gemini API, AI Studio, Android Studio, Google Antigravity, and the Gemini Enterprise Agent Platform; Gemini Spark has also begun switching to the new model for Google AI Pro and Ultra subscribers.
On the Artificial Analysis Intelligence Index, Gemini 3.7 Flash still trails Anthropic Claude Opus 5 which leads the overall ranking. Meanwhile, Google's flagship Gemini 3.5 Pro remains delayed due to coding capability gaps, with its originally planned August launch pushed back approximately two months.