Newsletter Subscribe
Enter your email address below and subscribe to our newsletter
[forminator_form id="25163"]

axiosmarktechpost+1axiosMeta released two AI products this week as it pushes to keep pace in an increasingly crowded model race and looks to accelerate its product pipeline following the resolution of major litigation.
On Wednesday, Meta released Muse Spark 1.3, an update the company says delivers improved performance in coding and agentic tasks. The model is rolling out in Muse Code and through Meta's API, priced the same as its predecessor — a move Meta AI chief Alexandr Wang called "aggressive".axios
"It's very competitive with frontier models," Wang told Axios. He tied the update to broader ambitions around personal agents, referencing goals CEO Mark Zuckerberg has outlined on earnings calls: "personal agents that can work 24/7 on your behalf and help you achieve your goals".axios
Wang also noted strong adoption of Meta's "contributor tier," which lowers coding product costs in exchange for allowing Meta to use developer work to improve its models. A "meaningful double digit" percentage of coders are choosing that option, he said.axios
The release lands amid a wave of competing announcements. Google on Wednesday debuted Gemini 3.8 Flash, Anthropic updated its Fable and Mythos models, and OpenAI said Tuesday it will soon release its Astra model.axios
A day earlier, Meta Superintelligence Labs released Muse Voice Transcribe, a real-time model that handles streaming speech recognition, speaker diarization for more than 20 simultaneous speakers, and endpointing in a single architecture. The model is available through the Meta Model API at $3.00 per 1,000 audio minutes and is integrated into Meta AI for Mac and Muse Code.marktechpost
Artificial Analysis placed Muse Voice Transcribe first on its streaming speech-to-text leaderboard at approximately 3.1% word error rate, ahead of Cartesia Ink-2 at 3.4% and ElevenLabs Scribe v2 Realtime at 3.6%. At its price point, Meta undercuts Cartesia at $4.00 and ElevenLabs at $6.50 per 1,000 minutes.fourweekmba+1
The model is API-only with no open weights released.marktechpost
The product launches come as Morgan Stanley analysts argue that Meta's roughly $18 billion settlement with 48 states over child-safety claims could free the company to ship AI products faster. "There are indeed signals indicating that Meta's product pipeline may begin to accelerate its release now that these legal obstacles have been cleared," the analysts wrote, drawing a parallel to Google's Alphabet Inc. product push after the DOJ decided not to force asset divestitures.moomoo+1
Not all analysts are as optimistic. Needham maintained a "Hold" rating, warning that the settlement costs compound Meta's capital expenditure pressures — projected at up to $145 billion in 2026 — and cautioning against "strategic dispersion" across too many AI initiatives at once.news.futunn+1