Liquid AI released LFM2.5-VL-3B-DSpark, an experimental 280M-parameter drafter it says speeds decoding of its LFM2.5-VL-3B vision-language model by up to 3.13x.
Speculative decoding is a common trick for speeding up text models; Liquid AI says the same recipe carries over to vision-language models, where image prefill usually dominates latency on smaller devices.
OpenAI's DevDay 2026 recap post says the event included more than 20 announcements spanning GPT-6 Astra, ChatGPT, Codex, APIs, security, and new builder tools.
OpenAI's @OpenAI account says GPT-6.1 Sol is a significant upgrade over GPT-6 Sol in coding, computer use and complex professional work, nearing GPT-6 Astra on some benchmarks at substantially lower cost.
AWS says Claude Sonnet 5.5 is now available on Amazon Bedrock and Claude Platform on AWS, pitched as a cheaper, faster Sonnet for scoped coding and enterprise tasks.
Mark Zuckerberg said on X that Meta's Muse personal agent now handles voice and real-time video, can work in the background during long conversations, and is coming to glasses for hands-free use.
Google's @GoogleAI account announced Gemini 4 Argon, a frontier model it says sustains deep reasoning across long-horizon workflows in software engineering and enterprise work like legal and finance.
Microsoft AI announced three new models on X — MAI-Transcribe-2-Streaming, MAI-Voice-2.1 and MAI-Voice-2.1-Flash — pitched at streaming transcription and faster turn-taking in voice agents.