Gemini 3.6 Flash: 17% fewer tokens, lower cost, and a Python cold start fix you didn't have to ask for

Short summary

Gemini 3.6 Flash cuts output tokens by 17% and lowers pricing to $1.50/$7.50 per 1M, with compounding savings across multi-step agent loops. Vercel AI Gateway adds the new Gemini models and Poolside's Laguna S 2.1 open-weight model with 1M context for large-codebase agents. Python cold starts drop by half with no code changes. Most updates are single-line config swaps worth deploying immediately.

• •Gemini 3.6 Flash: 17% fewer output tokens, $7.50/1M output, single-param migration
• •Vercel AI Gateway consolidates multi-provider routing with budget tracking and failover
• •Poolside Laguna S 2.1 open-weight model hits 78.5% SWE-bench with 1M context for repo-scale coding

Generated with AI, which can make mistakes.