Skip to main content
cto.new works with the best frontier LLMs for chat and code generation. Auto model automatically routes your task to the best frontier LLM, balancing performance, cost, speed, and capacity. Premium customers can choose which model to use for each task.
Be prepared to experiment with settings when choosing your own model. Performance may vary.
cto.new supports:
  • Claude Sonnet 4.5
  • Claude Opus 4.5
  • Claude Haiku 4.5
  • GPT-5.2
  • GPT-5.1 Codex
  • GPT-5.1 Codex Mini
  • GLM 4.7
  • Devstral 2
  • Minimax M2.1
  • Kimi K2 Thinking
  • Gemini 3 Pro Preview
  • Gemini 3 Flash Preview
  • Sonnet 4.5 - GPT-5.2 Alloy
  • Sonnet 4.5 - GPT-5.1 Codex Alloy
Model alloys combine two LLM models in the same agent thread. Our research shows it improves performance by several percentage points.