NAll news

All news

Tuesday89 items

  1. 36Kr · AIProduct release
    AI score82

    DeepSeek open-sources inference framework DeepInfer and Token codec component deepseek-recipe

    DeepSeek open-sources its self-developed inference framework DeepInfer (Apache 2.0) and the production-grade Token codec component deepseek-recipe (MIT), supporting DeepSeek V4.1 Flash in serving 155 trillion Tokens per day on a single cluster. Official online statistics show an average of 77K Tokens input per request, a time to first token of about 400 milliseconds, 230 Tokens output per second during the generation phase, and an average of 96% of input hitting the cache. DeepInfer currently supports V4.1 Flash; the open-source version provides only offline inference, and the KV cache management and load balancing code has not yet been fully open-sourced because it is tied to internal infrastructure.

    Same story · 29 reports

    #English text is being prepared.#Reasoning#English text is being prepared.

  8. SiliconANGLEProduct release
    AI score51

    Scaled Cognition Launches Customer Service Voice Model Voice

    Scaled Cognition has released Voice, a text-to-speech model that supports more than 25 languages and voice cloning, and can be accessed via API or deployed self-hosted. It focuses on accurately reading aloud details such as account numbers and addresses. Managed API pricing is $9 per million characters, while self-hosting costs $1 per million characters (excluding infrastructure). The company says that at a scale of 5 million characters, including infrastructure, the cost is about $13 per million characters, and it claims a speech generation latency of 70 milliseconds. Voice's Smart Pronunciation capability first determines the meaning of abbreviations and numbers before generating audio—for example, reading "Pay 1/2 by 1/2" as "pay half by January second"—and it can run without relying on APT or other large language models.

    #Model releases#Speech#English text is being prepared.

  11. Hugging FaceFirst-partyProduct release
    AI score62

    PyTorch 2.15 Brings FP8 and Faster MoE Decoding to MPS

    PyTorch 2.15 adds FP8 support to the MPS backend and makes MoE model decoding 5–6.5 times faster than 2.14: on an M3 Ultra with torch.compile, Qwen3.6-35B-A3B goes from 9.4 tokens/s to 61.3 tokens/s, and LFM2.5-8B-A1B goes from 23.6 to 123.6 tokens/s. For Wan2.2's tiled VAE video decoding, peak driver memory drops from about 270 GiB to 8 GiB, and the image corruption issue with Qwen-Image-2.1's VAE encoding on MPS is also fixed with a padding change, raising the encode-decode round-trip PSNR from 10.06 dB to 29.95 dB. This version requires macOS 15 or later and Python 3.11 or later.

    #Product updates#English text is being prepared.#Open-source ecosystem

  12. Hugging FaceFirst-partyProduct release
    AI score54

    Hugging Face Releases AnyLanguageModel 1.0, Unifying Local and Hosted Model Calls on Apple Platforms

    Hugging Face has released AnyLanguageModel 1.0, which lets developers use the same session code to call system models, open-weight models from llama.cpp, MLX, and Core ML, as well as multiple hosted service providers. This version extends tool calling and guided generation to all providers, supports attaching images in prompts, and adds support for the Anthropic Messages API, Gemini API, Ollama, Private Cloud Compute, and Hugging Face Inference Providers. The article also lists currently high-download models such as Gemma 4 E2B, LFM2.5 2.6B, and Qwen3.8 27B, along with their quantized sizes.

    #Product updates#Open-source ecosystem#English text is being prepared.

  15. Cursor BlogFirst-partyProduct release
    AI score67

    Cursor Releases Agent SDK, Enabling Code-Defined Agents with One-Click Deployment

    Cursor has released the Agent SDK, the framework it uses to build hosted agents such as Bugbot, Security Reviewer, and Rollouts. Developers can define agents in code and deploy them to Cursor infrastructure with a single command, agent-sdk deploy. Agents are organized as small software projects, with prompts written in Markdown and tools defined as TypeScript functions. They can respond to PR events from GitHub, GitLab, and Bitbucket, and include built-in evals and event replay. They can also be packaged to run as a single Node process on your own infrastructure.

    #Product updates#Agents#English text is being prepared.

  16. ayi_ai_notesProduct release
    AI score69

    English translation is being prepared.

    Quoted postDeepSeek · @deepseek_ai

    🚀 Introducing DeepInfer, our in-house inference framework, now fully open source! Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable. 🔗 https://t.co/JGAik1kgk0 1/4

    Same story · 29 reports

    #English text is being prepared.#Reasoning#English text is being prepared.

  17. Shao MengProduct release
    AI score84

    Alibaba Tongyi MAI-UI team open-sources general computer-use agent tech stack OpenGCUA

    Alibaba Tongyi MAI-UI team open-sources the complete general computer-use agent tech stack OpenGCUA, including harness, cross-device SDK, native macOS app, evaluation trajectories, and reference agent implementation. Its core design is that code execution and decision-making shift from the harness to the model; the model takes screenshots on demand, writes Python programs with loops and conditional judgments to orchestrate GUI operations, API calls, and shell commands, and maintains its own todo list. In evaluations of 8 frontier models on OSWorld-v2, the average partial score rose from 45.3 to 64.9, and the average number of steps per task decreased by 53%; in real-machine macOS tests, under the same harness, Qwen3.8-Max's inference token consumption was 19.4 times that of GPT-6 Astra, completing 3 of 4 tasks.

    #English text is being prepared.#Agents#English text is being prepared.

  18. Larsen CundricProduct release
    AI score58

    Grok Bot launches @bot tagging on X

    Larsen Cundric announced that @bot tagging on X is now open, and users whose X and Grok accounts are already connected to the Bot can use it right away. Just reply to any post with @bot plus an instruction, for example adding it to a Notion reading list, setting a reminder, summarizing a thread, or drafting a reply, and the result lands in the corresponding Grok Bot.

    Quoted postLarsen Cundric · @larsencc

    NEW: @bot tagging just went live. Reply to any post: > @bot add this to my Notion reading list > @bot remind me to read this tomorrow > @bot summarize this thread > @bot draft a reply to this Works in replies, posts, quote and lands right in your Grok Bot. What’s the first thing you’re handing it?

    #Product updates#English text is being prepared.#xAI

  19. Xie SainingProduct release
    AI score72

    Dream Labs releases robot foundation model Dream-1, starting at $3 per hour

    Dream Labs has released the robot foundation model Dream-1, saying it can control dual-arm robots it has never seen before, complete tasks it was not trained on, and enter places it has never been, starting at $3 per hour. The Dream API opened to everyone the same day. Xie Saining reposted the news and congratulated the team.

    Quoted postDream Labs · @dreamlabs_ai

    Today, Dream Labs is coming out of stealth with Dream-1, a robot foundation model that controls bimanual robots it has never seen, on tasks it was never shown, in places it has never been, from $3/hour. Dream API is live for everyone, today. 🧵 https://t.co/MB1snhUOAJ

    #English text is being prepared.#Product updates#Embodied AI