Announcing MCP Toolbox Java SDK v1.0: Agentic data access for the enterprise
cloud.google.com·6d ago
Cloud & Infrastructure
The New Stack·Aug 13·5 min read
The AI model that just scored 65% on DeepSWE isn’t the one Google promised.
TL;DR
Google has launched the Gemini 3.7 Flash model, which improves coding capabilities and workflow management, achieving a 65.3% score on DeepSWE v1.1. This model is priced at half the cost of its predecessor, Gemini 3.6 Flash, until a price increase in 2027.
✦ Why It Matters
Evaluate the Gemini 3.7 Flash model for your coding and workflow automation needs before the price increase.
Key Takeaways
- Test Gemini 3.7 Flash at the current discounted rate before the price increase.
- Adjust API settings according to Google's migration guide for smooth transition.
- Consider the thinking level settings to optimize performance for specific tasks.
- Monitor automation performance to ensure reliability in production environments.
How It Works
Gemini 3.7 Flash utilizes three thinking levels—low, medium, and high—to optimize processing based on task complexity. Low is for latency-sensitive tasks, medium balances speed and reasoning, while high allows for deeper analysis in challenging scenarios.