
022
DeepSeek Retires Its 1.6T Model for a 552B One
No.22 · Friday, September 11, 2026 · Today in AI
Terms in this issue
In this issue
- 01DeepSeek Retires Its 1.6T Model for a 552B OneRequests to V4-Pro, released only a month ago, move to the new model on September 14. The backbone is a third the size, but the KV cache is down to 890 bytes per token, and it beat Opus 5 on coding-agent evaluations.

- 02Eighteen models that run on a five-year-old phoneA new European AI lab has built one model per task. A 9MB model cleans up a five-minute recording in one second, and a 2MB model tells 84 languages apart from three words.

- 03OpenAI mobilized 250 people to hunt its own vulnerabilitiesThe company declared an internal code red and swept more than 100 service areas. It has published the whole process as a reference architecture and running order called Defense Factory.

- 04Apple Unveils the Folding iPhone Duo and 2nm A20 ProThe Neural Engine now comes as two 16-core blocks, doubling on-device AI compute. iOS 27 with Siri AI arrives September 14; iPhone Duo ships October 23 starting at $1,999.

- 05OpenAI releases an agent that queries company data in plain languageIt connects straight to Snowflake or BigQuery, investigates what changed and builds the dashboard for you. No query language required, and the row and column permissions on your account still hold.

- 06Tencent Releases an Open-Source Model That Generates and Edits SpeechGive it a written instruction and it produces a voice, then changes the emotion and delivery. The code and weights are out under an MIT license, and a distilled version that answers in four steps shipped alongside it.

Subscribe
Every weekday at 6 AM KST: the AI news that matters, plus the Hello, Alien letter and a term of the day — all in one issue. Weekends off.
By subscribing you agree to our Terms of use and Privacy policy.
Free · unsubscribe anytimeSee you in the next issue
