one year on
DeepSeek releases R1-0528 reasoning model, adds JSON output and function calling
The update posts improved benchmark performance and adds JSON output and function calling.
DeepSeek releases R1-0528, an updated version of its open-weight reasoning model, on Hugging Face today. The company’s API docs highlight improved benchmark performance, enhanced front-end capabilities, reduced hallucinations, and new support for JSON output and function calling.
DeepSeek says the update improves benchmark performance and is available on Hugging Face and via the DeepSeek chat interface.
The record
One year later — open only if you can handle spoilers
R1-0528 became a widely used baseline for reasoning model comparisons throughout mid-2025, especially among open-weight enthusiasts.
The Weekly Replay · free by email
This week, one year ago — every Sunday.
One email each Sunday: the week's replayed AI news, with the one-year-later annotations included. Written like it's breaking — dated like it isn't.
Free · double opt-in · unsubscribe anytime · privacy