Today we are releasing JST-2, the world's first full-stack physics-native world foundation model. Many world models learn in pixel/video space. We went back to first principles and built the entire model stack in 3D. Generate a 3D scene in seconds and expand it in real time. Create characters with motion built in, all through one general architecture.
#Product updates#Video#Multimodal
Perplexity Computer can now run fully local on your Mac. Work with your native Mac apps and local files, and run tasks without sending sensitive data to the cloud. The model and the agent both run on your Mac. Available now in the Perplexity Mac app. https://t.co/tnvCgCXD7v
#Product updates#English text is being prepared.#English text is being prepared.
#Product updates#Agents#RAG
#Product updates#English text is being prepared.#Agents
Google announced that Gemini is now available globally on Android Auto and in cars with Google built in, and has begun rolling out Personal Intelligence, which, with authorization, can connect to Workspace (including Gmail and Google Calendar), Google Photos, Google Maps, YouTube, and Search to answer questions such as flight gate numbers, restaurants along the route, and locations in photos. Personal Intelligence begins rolling out on Android Auto today, with cars with Google built in to follow over the coming months. It is currently available to eligible personal Google accounts for users aged 18 and over, excludes Workspace business, enterprise, and education users, and is not available in the European Economic Area, Switzerland, the United Kingdom, or Nigeria.
#Product updates#Google#English text is being prepared.
#Product updates#OpenAI
#English text is being prepared.#Open-source ecosystem#Google
#Product updates#Google#Speech
#Product updates#Agents#English text is being prepared.
The SGLang team, in collaboration with Google and Meta, has open-sourced SGLang-TorchTPU, a PyTorch-based native backend that brings the full SGLang serving stack to Google Cloud TPU and is already in large-scale production use at Meta. Initial support covers models including Muse Spark Open, GLM-5.3, Kimi K3, DeepSeek V4, Qwen3.8, and Gemma 4, with features such as speculative decoding, KV cache offloading, PD disaggregation, and FP8 KV cache. The team also used TorchTitan and Miles to complete GRPO training of Qwen3-4B on a single 4-chip TPU v7x, with the reward curve matching a 6× H200 baseline within random fluctuation.
#English text is being prepared.#English text is being prepared.#Open-source ecosystem
#Industry news#Agents#Safety / alignment
DeepSeek has open-sourced its internal large-scale cloud service inference framework DeepInfer and the production-grade token codec component deepseek-recipe. The former supports distributed inference, speculative decoding, SWA Bounded Replay, and structured output, and can run on NVIDIA and Huawei Ascend platforms. DeepSeek says the high-throughput, low-latency inference service of DeepSeek V4.1 Flash is based on this framework, with a single cluster serving 155 trillion Tokens/day in throughput, at 38 US dollars per card per hour calculated at off-peak prices. The processing logic of deepseek-recipe is fully aligned with the official DeepSeek API. The team also explained that the KVCache management and load balancing components are deeply coupled with internal infrastructure, so the relevant code has not yet been fully open-sourced.
#English text is being prepared.#Reasoning#English text is being prepared.
#Product updates#English text is being prepared.#Agents
#Product updates#Agents#English text is being prepared.
Gemini's Personal Intelligence feature is beginning to roll out to Android Auto; once enabled, it can connect to user data in Gmail, Google Calendar, Google Photos and Google Search to provide more personalized answers. The feature is entirely opt-in in Gemini settings, and in the coming months it will also arrive on built-in Google cars that run Google's in-car system directly. Karan Daryanani, a senior Google product manager, said the feature's two core capabilities are extracting specific details from emails, photos or saved places in Google Maps, and reasoning across complex sources.
#Product updates#Google#Agents
#Product updates#Microsoft#Agents
Beam is a new Mac app that can pull a single window from another Mac onto your own desktop and use it like a local app. Its goal is to make it easy to access a Mac Mini that sits in a corner running AI agents around the clock. The app was developed by Hiten Shah and Sam Asante under the company name Noats and launched on Product Hunt on Tuesday. Both Macs must be Apple Silicon, run macOS Sonoma or later, and have Beam installed; they can discover each other over the same Wi-Fi or via Tailscale. Hosting a Mac and joining via a share link are free; using your own Mac is free for 20 minutes per day, after which it costs $12 per month, $84 per year, or a $199 lifetime license.
#Product updates#English text is being prepared.#Agents
DeepSeek has open-sourced DeepInfer, its in-house inference framework used to serve its models in production. Released under the Apache 2.0 license, it lets developers run DeepSeek-V4.1-Flash on their own infrastructure, with support for subsequent models planned. The framework includes optimized compute kernels and communication libraries such as FlashMLA, DeepGEMM, and DeepEP, supports speculative decoding, asynchronous scheduling, and structured output, and also open-sources the deepseek-recipe library (available in Rust and Python versions) for handling request and response transformations. On the hardware side, it supports Nvidia SM100 (Blackwell) architecture and the Huawei Ascend 950 NPU series from day one; older architectures such as Hopper are not listed. Batch-invariant inference and deterministic sampling are available only on Nvidia hardware, and the published performance data is also based on Nvidia GPUs. The current version supports offline inference but lacks complete production serving components such as KV cache management, cross-machine inference stage separation, and replica load balancing.
#English text is being prepared.#Reasoning#English text is being prepared.
ElevenLabs has released Eleven Realtime, offering two hosted voice agent models, Mini and Max, which integrate speech recognition, reasoning, turn-taking, and speech generation, but developers cannot individually select or configure the underlying models. Data from external testing organization Artificial Analysis shows that Max scored 98.8% on Big Bench Audio, while Mini scored 75.9%; for tool-calling task completion, Max reached 87.1% and Mini 60.3%, while Mini slightly outperformed Max on Full-Duplex-Bench with 95.8% versus 94.1%. The interface follows the OpenAI Realtime API protocol and WebSocket transport, but only supports WebSocket; WebRTC applications must first switch transport methods and handle the capture and playback of 24kHz PCM16 audio themselves.
#Model releases#Speech#Agents
#English text is being prepared.#English text is being prepared.#English text is being prepared.