OpenAI's Jalapeño chip posts first measured numbers — 1.5–1.9× more AI work per watt and 1.7–3.6× lower end-to-end latency
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 ships V4-Flash-Vision-Exp — its first multimodal model, matching V4-Flash on text, adding image input
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` — an MIT agent harness where model adapter, tools, and the agent loop itself are all replaceable plugins
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.