News2026-10-22 14:00 UTC

Recursive launches Flint local model

40B-parameter mixture‑of‑experts model activates 2.5B per token, runs 125‑150 tokens/sec on MacBook Pro and NVIDIA RTX Spark, enables fast agentic coding locally without sending data to remote servers.

More News

10-22 14:00 UTCMatt Crosby: Bitcoin Primed for Big Move Higher
10-22 14:00 UTCAnthropic unveils Claude Duet
10-22 14:00 UTCOpenAI models to appear on Google Cloud
10-22 14:00 UTCAnthropic releases Claude API version 2026‑10‑21
10-22 14:02 UTCTokenized equities market sees rotating leaders as volume hits $25 B