OpenAI ships Decisions API for 10x faster bounded reasoning as open foundation models expand into DNA and Chinese infrastructure
The past 24 hours delivered critical cost and architecture advancements for builders: OpenAI debuted an ultra-fast, zero-output-cost routing and classification API, Hugging Face open-sourced a 14B model rivaling frontier systems on biological sequence tasks, and iFlytek launched a 1M-token MoE built entirely on domestic silicon.
OpenAI Launches Low-Cost Decisions API Built on GPT-6 Luna
SourceDramatically cuts costs and latency for high-throughput pipeline tasks like agent intent routing and data classification with $0.10/M token pricing and no output fees.
Hugging Face Drops Carbon-X, an Open 14B DNA Model
SourceEnables biotech founders to run specialized sequence modeling locally at roughly 1,000x lower inference cost than proprietary frontier LLMs.
iFlytek Open-Sources Spark-X2.6 on Domestic Infrastructure
SourceOffers a 1M context window 293B-parameter MoE (30B active) that operates independently of Western hardware restrictions for enterprise agents.
Meta Unveils Muse Spark Open 816B Parameter Model
SourceExpands open-weights capabilities to massive scale, allowing teams building heavy reasoning workflows to deploy without vendor lock-in.
Get tomorrow's digest the moment it ships.