Google Gemini 3.7 Flash Launches at Half Price as 3.5 Pro Delay Drags On
Google launches Gemini 3.7 Flash with stronger coding scores and half-price introductory rates, while its delayed Gemini 3.5 Pro remains missing in 2026.
Summary
On August 13, 2026, Google began rolling out Gemini 3.7 Flash, replacing 3.6 Flash just three weeks after its release instead of delivering the long-awaited flagship Gemini 3.5 Pro. Google credits core optimizations and developer feedback for improved coding and agentic performance. Senior Director Tulsee Doshi cited FrontierCode 1.1 Main rising from 34.4 percent to 43.6 percent, DeepSWE v1.1 from 49 percent to 65.3 percent, and WebDev Arena from 1,538 to 1,588. GDP.pdf, which tests complex-document processing, improved from 22 percent to 34 percent, while AutomationBench, covering common business workflows, climbed from 17 percent to 30.4 percent.
The introductory price through year-end is $0.75 per 1 million input tokens and $3.75 per 1 million output tokens, half 3.6 Flash’s rates but above OpenAI’s reduced GPT 5.6 Luna prices of $0.20 and $1.20, respectively. Gemini 3.7 Flash is live in the Gemini API, AI Studio, and Gemini Enterprise; consumers receive it only through the Gemini Spark agent with AI Pro or Ultra, while the regular chatbot remains on 3.6 Flash. After Google gained ground during 2024 and 2025, progress appears slower in 2026: Gemini 3.5 Pro, promised at I/O in May for June, remains unreleased. Ars Technica suggests multiple Flash versions approaching 4.0 may avoid comparing 3.5 Pro with newer OpenAI and Anthropic models; reports say Gemini coding has lagged rival labs as Google loses AI talent.
Positives
- FrontierCode 1.1 Main increased from 34.4 percent to 43.6 percent, indicating stronger coding performance than Gemini 3.6 Flash.
- DeepSWE v1.1 rose from 49 percent to 65.3 percent, while WebDev Arena advanced from 1,538 to 1,588.
- GDP.pdf improved from 22 percent to 34 percent, and AutomationBench climbed from 17 percent to 30.4 percent.
- $0.75 input and $3.75 output pricing per 1 million tokens halves Gemini 3.6 Flash’s rates through year-end.
- Gemini API, AI Studio, and Gemini Enterprise customers can begin using Gemini 3.7 Flash immediately.
Risks & concerns
- Gemini 3.5 Pro missed its promised June launch after Google announced the flagship model at I/O in May.
- OpenAI’s GPT 5.6 Luna remains cheaper at $0.20 per 1 million input tokens and $1.20 per 1 million output tokens.
- Regular Gemini chatbot users remain on 3.6 Flash because 3.7 Flash consumer access requires AI Pro or Ultra and the Spark agent.
- Reports say Gemini’s coding capabilities have fallen behind recent rival advances while Google experiences an exodus of AI talent.
- Gemini 3.7 Flash replaced 3.6 after three weeks, prompting questions about whether its benchmark gains justify another model release.