DeepSeek ships V4.1-Flash: asymmetric MoE, 1/4 the KV cache
552B MoE, native vision, 8B/16B active params, MIT weights. The KV-cache compression DeepSeek wants agents to live with.
OpenAI Jalapeño chip posts first benchmarks
OpenAI published first-party benchmark numbers for Jalapeño: 1.5–1.9× more AI work per watt, 1.7–3.6× lower end-to-end latency. Vendor-tested, not independent.
DeepSeek V4-Flash-Vision-Exp: first multimodal model
DeepSeek's experimental vision model matches V4-Flash on text tasks and adds image input at the same pricing; new Files API lets you upload once and reuse across requests.
DeepSeek ships `dsh`: MIT agent harness
DeepSeek's dsh agent harness (MIT) hit ~165K GitHub stars in six days; every part — model adapter, tools, agent loop — is a replaceable Cordis plugin. Developer preview.