Google Gemini 3.7 Flash Boasts Speed Gains but Price Limitations

Google recently launched Gemini 3.7 Flash, its latest high-speed model for coding, agents, and knowledge work. This update follows just three weeks after the previous version, Gemini 3.6 Flash, was released.

The new model is designed to follow instructions more reliably and recover from roadblocks across multi-step tasks. It has been benchmarked across various areas, including coding, web development, document analysis, and automation, with consistent improvement over its predecessor.

Google's introductory rate for Gemini 3.7 Flash is $0.75 per million input tokens and $3.75 per million output tokens, which is half the original price of Gemini 3.6 Flash. However, this discounted price expires on December 31, 2026, and will increase to $1.50 for input and $7.50 for output in January 2027.

While Gemini 3.7 Flash performs well across various benchmarks, it's not the best or cheapest model for every workload. Other models like GPT-5.6 Luna and DeepSeek V4 Flash currently undercut its price. Additionally, Google's own comparison shows that other models lead in specific tasks.