Factory releases Droid 1.0: Code Mode enabled by default, significantly reducing token consumption.

Beating AI News Flash: Factory has released Coding Agent runtime framework Droid 1.0, introducing programmatic tool calling (Code Mode) and setting it as the default execution method. AI can first write JavaScript scripts and then use the scripts to call tools hundreds of times in succession, without needing to re-request the large model every few steps, thereby reducing the number of model calls and Token consumption.

For example, to migrate and verify 151 Linear tickets, the traditional approach is estimated to require 36 model requests and 3.1 million input Tokens. After switching to Code Mode, only 2 requests and about 61,000 Tokens are actually needed, reducing input Tokens by about 98%. In broader tests, task costs using Claude Opus 5.5 were reduced by about 15%.

Factory also optimized tool-calling strategies, caching, and task execution mechanisms for different models. In the official Terminal-Bench 4 self-test, using the same GPT-6.1 Sol model, Droid scored 3.5 points higher than Codex, with successful task costs reduced by 12%, but time consumption increased by 10%.

Code Mode is not a new technology. Cloudflare and Anthropic had previously disclosed similar solutions, and Pi, OpenCode 2.0, and DeepSeek Harness also already support it.