Models

Google launches Gemini 3.6 Flash with 17% cost cut

Google has launched Gemini 3.6 Flash and 3.5 Flash-Lite, cutting output costs by 17 percent while boosting coding benchmarks to offer developers more efficient and affordable AI models.

AlphaSignal31 Jul 2026Models
Image: AlphaSignal

Google has updated its AI lineup with the release of Gemini 3.6 Flash, which replaces Gemini 3.5 Flash as the default workhorse model. The new model reduces output token usage by up to 17 percent, resulting in lower costs for developers. Specifically, Gemini 3.6 Flash is priced at $1.50 per million input tokens and $7.50 per million output tokens, representing a drop from the $9.00 output rate of its predecessor. It retains a 1 million-token context window and features an updated knowledge cutoff of March 2026.

For practitioners, these efficiency gains are paired with notable performance improvements. On the DeepSWE coding benchmark, Gemini 3.6 Flash scores 49 percent compared to the 37 percent achieved by Gemini 3.5 Flash. It also shows progress on the MLE Bench for machine learning research, scoring 63.9 percent against the previous 49.7 percent. These metrics translate to higher precision in real-world software engineering, with fewer unwanted code edits and reduced execution loops during development.

Alongside the main update, Google introduced Gemini 3.5 Flash-Lite for high-throughput, low-latency tasks like document processing and agentic search. Operating at 350 tokens per second, this model is priced at $0.30 per million input tokens and $2.50 per million output tokens. Despite its low cost, it outperforms the older Gemini 3 Flash on key evaluations, scoring 54.2 percent versus 49.6 percent on SWE-Bench Pro, and 74.0 percent versus 65.1 percent on OSWorld-Verified.

Google also rolled out several ecosystem updates, though Gemini 3.5 Pro remains in partner testing. The Gemini desktop app for macOS now features system-wide voice control via the function key, enabling screen-aware rewriting and image editing across open windows. Additionally, the Gemini Spark agent has expanded to over 160 new countries for paid subscribers, integrating with Chrome to manage web tasks. Finally, the Nano Banana tool now offers free personalized image generation for eligible users in the United States, and new integrations have launched for Dropbox, Viator, and Zillow.

This is our own summary of reporting by AlphaSignal

More in Models