Google rolls out Gemini 3.7 Flash, a new model that improves coding and agentic performance while offering a lower introductory price.

Google unveiled Gemini 3.7 Flash on Thursday, marking the company’s fastest succession of major AI model updates in a single month. The new iteration builds on the capabilities of its predecessor, Gemini 3.5, by delivering faster code‑generation, more reliable agentic reasoning, and a reduced introductory pricing tier for enterprise customers.

Key upgrades in Gemini 3.7 Flash

The flash‑focused release trims latency by roughly 15% across benchmarked coding tasks, allowing developers to receive completions in near‑real‑time. Under the hood, Google added a larger context window—up to 128 k tokens—so that the model can handle longer codebases without losing track of earlier references.

Agentic performance also received a boost, with Gemini 3.7 Flash better at maintaining goal‑oriented dialogue over extended interactions. Early testers report fewer “hallucinations” when the model is asked to orchestrate multi‑step workflows, a common pain point for earlier versions.

Pricing and availability

Google introduced a new entry‑level plan priced at $0.009 per 1,000 tokens, undercutting the standard Gemini 3.5 offering. The tier is aimed at startups and hobbyist developers who need high‑quality code suggestions without the enterprise‑grade cost.

How the update fits into Google’s roadmap

The rapid rollout follows Google’s stated goal of delivering “continuous improvement” to its generative AI portfolio. By spacing releases just three weeks apart, the company signals confidence in its development pipeline and a desire to stay ahead of competing models from OpenAI and Anthropic.

  • Improved token efficiency for longer prompts
  • Enhanced code‑completion accuracy across 12 programming languages
  • Reduced hallucination rates in multi‑turn agentic tasks
  • Lower entry‑level pricing for broader adoption
We’re listening to the developer community and iterating faster than ever before, delivering tangible performance gains at a more accessible price point, said Google’s vice‑president of AI products in a launch briefing.

Analysts note that the incremental nature of Gemini 3.7 Flash could pressure rivals to accelerate their own update cycles, especially as enterprises weigh cost against capability in a crowded AI market.

Ars Technica coverage of Google’s Gemini 3.7 Flash launch