Google has released Gemini 3.6 Flash, an updated version of its AI model, which is faster and more cost-effective than its predecessor. The new model is designed to improve efficiency, particularly in agentic workflows, where it can complete tasks more accurately with fewer steps and tokens. According to the company, these changes were made in response to user feedback on the earlier 3.5 Flash version. The model now supports computer use as a standard feature in the Gemini API, with a modest improvement in performance compared to the previous version. Source: arstechnica
In the DeepSWE test for coding, Gemini 3.6 Flash achieved a score of 49 percent, compared to 37 percent for 3.5 Flash. The OSWorld test for computer use also showed a slight improvement, with the new model reaching 83 percent from 78.4 percent. Additionally, Gemini 3.6 Flash has a lower API cost, at $1.50 per 1 million input tokens and $7.50 per 1 million output tokens, compared to $1.50 and $9 for 3.5 Flash. Token efficiency has been significantly improved, with the model using about 17 percent fewer tokens. Source: arstechnica
Google also released Gemini 3.5 Flash Lite and 3.5 Flash Cyber. The new Flash Lite is the most efficient modern AI model from Google, achieving 350 tokens per second. It is positioned as an affordable option for scaling agentic systems. The model’s pricing is set at $0.30 per 1 million input tokens and $2.50 per 1 million output tokens, slightly higher than the previous 3.1 Flash Lite. Source: arstechnica
