Gemini 3.7 Flash outpaces 3.6
Google’s Gemini 3.7 Flash is a generally available workhorse model for coding, agents, multimodal reasoning, and knowledge work. It retains a 1M-token context window while improving substantially over Gemini 3.6 Flash on agentic coding and automation benchmarks.
Gemini 3.7 Flash looks like a meaningful efficiency win, especially for developers building fast, tool-using agents, though its gains are not universal across every evaluation.
- –Terminal-Bench 2.1 rises from 78.0% to 85.8%, pointing to stronger real-world coding-agent performance
- –Code Arena Elo improves from 1538 to 1588, suggesting better web-development output
- –Native function calling, search, and computer-use support make it practical for multimodal agent workflows
- –The 1M-token context and 64k output limit support large repositories, documents, and long-running tasks
- –Introductory pricing is $0.75 per million input tokens and $3.75 per million output tokens through December 31, 2026
DISCOVERED
51m ago
2026-08-13
PUBLISHED
59m ago
2026-08-13
RELEVANCE
AUTHOR
browser_use