GPT-5.5 Instant is now the default — what API developers should know

On May 5, OpenAI updated ChatGPT’s default model to GPT-5.5 Instant — available to all users and accessible in the API as chat-latest. This is the fastest GPT yet, optimized for agent workflows.

What changed

• • Speed: 2-3x faster than GPT-5 for most prompts
• • Cost: Lower token pricing → agents can make more API calls per budget
• • Tool use: Improved function calling accuracy
• • Default: All ChatGPT users get it automatically

Impact on apitree

Faster models mean agents make more API calls per session. An agent that used to make 3 API calls now makes 8 — because it can iterate faster. This is good for apitree: more calls = more usage data = better recommendations.

If you’re using apitree with OpenAI agents, no changes needed. The proxy handles any model — GPT-5.5, Claude, Gemini. apitree is model-agnostic.

Source: LLM Stats