←company
← All topics

company

English text is being prepared.

Public reading groups
509
Last 30 days
170
Directory status
Legacy topic

LatestOpenAI releases the GPT-Rosalind-6 system card addendum, offering restricted access for biological research

Latest reading groups

Duplicate reports of the same fact are grouped

  1. OpenAI: Deployment Safety and System CardFirst-partyModel releases
    AI score72

    OpenAI releases the GPT-Rosalind-6 system card addendum, offering restricted access for biological research

    OpenAI releases the GPT-Rosalind-6 system card addendum. This model uses the same set of weights as GPT-6 Astra, with no new training data or post-training added, and is provided only through trusted access to vetted research institutions, enterprises, and government partners. At the highest reasoning budget, MedChemBench v7 averages 49.6% (GPT-Rosalind-5.5: 28.9%), GeneBench Pro v13 single-attempt success rate is 37.7% (previous generation: 16.7%), and LifeSciBench Gold average score is 61.3% (previous generation: 44.0%). Under the Preparedness Framework, this release is rated High capability in the biological and chemical domains, below the Critical threshold; rated Critical in the cybersecurity domain; and below High in AI self-improvement.

    Why read thisThe system card discloses specific scores and threshold gaps for biological-domain capability evaluations, which can be used to understand the openness boundary of frontier models in dual-use research.

  1. UnslothFirst-partyProduct release
    AI score72

    Unsloth releases GGUF version of DeepSeek-V4.1-Flash-1017

    Unsloth releases a new GGUF version of DeepSeek-V4.1-Flash-1017, claiming it ranks highest in benchmark tests. It says Dynamic V3.1 leads on almost all benchmarks including GPQA Diamond, AIME25, LiveCodeBench v6, MMLU-Pro, and Aider, requires more than 64GB of memory to run, and includes links to the GGUF and the blog.

    Why read thisUnsloth has released a GGUF quantized version of DeepSeek-V4.1-Flash-1017, along with a 64GB memory requirement, making it easier for local deployers to evaluate.

  2. CursorFirst-party
    AI score60

    Tampered Claude Code extension appears in Cursor extension marketplace, affecting up to 5,000 users

    Cursor said that a tampered Claude Code VS Code extension appeared in its IDE extension marketplace, affecting up to 5,000 users; the malicious version was not developed by Anthropic. Cursor took down the extension marketplace the same day, worked with relevant service providers to remove the malware, and after confirming there was no further risk, restored it online and required all users to upgrade to Claude Code for VS Code version 2.1.314.

    Why read thisCursor officially disclosed that a tampered Claude Code extension appeared in its extension marketplace, and provided the scope of impact and the handling process, which can serve as a reference for those concerned about IDE plugin supply chain risks.

  3. Cursor BlogFirst-party
    AI score75

    Malicious version of Anthropic.claude-code 2.1.313 appears in the Cursor plugin marketplace; official security advisory released

    Cursor discovered a maliciously tampered version of Anthropic.claude-code 2.1.313 in the VS Code extension marketplace. This version was not developed by Anthropic and affects at most 5,000 users. The malicious code collects host and user information, attempts to write the attacker's SSH public key and establish a reverse SSH tunnel, but due to permission restrictions this step fails on most macOS and Windows hosts. Cursor has removed the version, rebuilt the marketplace and shut down the attacker's server, and also provided a command to detect the svc-telemetry public key along with cleanup scripts for macOS and Windows.

    Why read thisCursor officially disclosed that a malicious version was planted in the plugin marketplace and provided detection commands and cleanup scripts so developers using the extension can check their systems step by step.

  4. Waymo：BlogFirst-party
    AI score60

    Toyota and Waymo Deepen Autonomous Driving Cooperation, Testing bZ4X Autonomous Fleet in Japan in 2027

    Toyota and Waymo announced that their autonomous driving cooperation has entered its next phase, with initial testing and validation of a Toyota bZ4X autonomous fleet to begin in Japan in 2027, followed by expansion to the UK and other Waymo operating markets over the following years. The vehicles will be produced at Toyota's Takaoka Plant in Toyota City, Aichi Prefecture, with everything from complete vehicle assembly to the integration of Waymo's autonomous driving hardware completed at the same plant. Waymo co-CEO Dmitri Dolgov said that in Waymo's operating areas, its rate of serious injury crashes is up to 20 times lower than that of human drivers.

    Why read thisThe original text provides the timeline and production arrangements for the Toyota-Waymo cooperation, allowing readers to gauge the pace of robotaxi deployment in Japan.

  5. ARC PrizeFirst-partyAnalysis
    AI score62

    ARC Prize publishes Meta Muse Spark Open's evaluation results on ARC-AGI

    ARC Prize published the validation results of Meta's Muse Spark Open on ARC-AGI: ARC-AGI-1 at 89.0% with $0.29 per task, ARC-AGI-2 at 57.5% with $0.73 per task, and ARC-AGI-3 at 0.1% under the standard harness with $3.9K. ARC Prize states that its v1/2 scores are close to those of Haiku 5.5 and GPT-6 Luna, but the cost is about 9 times that of the latter; the model is priced on OpenRouter at $1.25 per million input tokens and $4.25 per million output tokens. As the reasoning tier increases, ARC-AGI-2 rises from 29.4% at medium (averaging 55k tokens/task) to 57.2% at xhigh (180k tokens/task).

    Why read thisARC Prize gives Meta's new model's scores on three versions of ARC-AGI along with per-task cost, directly comparable with competitors scoring the same.

  6. PromptArmorFirst-partyAnalysis
    AI score60

    PromptArmor testing found that DeepSeek, Kimi, GLM, and Qwen are 88% more susceptible to cyberattacks than…

    PromptArmor tested its most widely circulated attack samples across multiple models and found that DeepSeek, Kimi, GLM, and Qwen models are 88% more susceptible to cyberattacks than GPT, Claude, and Gemini models. The tests covered multiple cyber risk dimensions, including susceptibility to phishing, data exfiltration, and running malware, and the authors described the differences as significant.

    Why read thisPromptArmor ran a cross-model comparison using its own widely circulated attack samples, revealing differences between Chinese and foreign models in phishing, data exfiltration, and running malware.

  7. Kimi.aiFirst-partyProduct release
    AI score64

    Kimi Launches 3D Make, Enabling Modeling and Game Development via Blender and CAD Tools

    Kimi has launched 3D Make, which can connect to tools such as Blender and CAD to create models, assemble scenes, and develop games. The feature is now available on the Kimi official website and Kimi Work, and the official team has also released 5 usage tips and step-by-step instructions.

    Why read thisThe official announcement explains Kimi's modeling, scene-building, and game development capabilities after integrating with Blender and CAD tools, and provides access points.

  8. Tencent HunyuanFirst-partyResearch
    AI score62

    Tencent Hunyuan Releases Hy-Nomas Technical Report and Open-Sources Hy-Nomas-Bench

    Tencent Hunyuan has released the technical report for Hy-Nomas, an interactive video world model, and has open-sourced Hy-Nomas-Bench for reproducible research. Officially, Hy-Nomas was also used to reproduce Genie 4's examples and a comparison video was released, along with an online demo that requires no waitlist. Earlier introductions stated that Hy-Nomas supports free interaction with text prompts and camera trajectories, long-horizon generation with world memory, lightweight LoRA adaptation for fields such as robotics, and optimization for inference on consumer-grade GPUs.

    Why read thisThe original text provides the technical report, the open-source benchmark, and a comparison video with Genie 4, so readers can judge the reproducibility of this world model.

  9. CognitionFirst-party
    AI score60

    Cognition Announces Acquisition of Corridor

    Cognition announced the acquisition of Corridor and called securing global software security the most important issue in the technology field today. The tweet also includes links to announcements posted by jackhcable and AshwinRamaswami.

    Why read thisCognition officially announced the acquisition of Corridor, positioning software security as the most important issue in the current technology field.

  10. UnslothFirst-partyProduct release
    AI score81

    Unsloth Releases 1-bit GGUF, Mistral Le Chonk Can Run Locally on Mac Studio

    Unsloth has released a 1-bit GGUF quantized version of Mistral Le Chonk, enabling this 1-trillion-parameter model to run locally on a Mac Studio, and claims that it still retains 91% top-1 accuracy after quantization. Unsloth also provided the GGUF download link and usage guide, and called Le Chonk the strongest Western open-source model currently available.

    Why read thisUnsloth reports that the 1-bit GGUF quantization retains 91% top-1 accuracy, allowing readers to judge the feasible boundaries of running trillion-parameter models on local devices.

  11. Factory Research / ProductFirst-partyProduct release
    AI score66

    Factory releases the Droid 1.0 agent framework, with programmatic tool calling enabled by default

    Factory releases Droid 1.0, the largest update yet to its agent framework, enabling programmatic tool calling (PTC) by default, which lets the model write JavaScript scripts to execute hundreds of tool calls within a single request in a sandbox. On Terminal-Bench 4, when paired with GPT-6.1 Sol, Droid scores 3.5 points higher than Codex at 12% lower cost; when paired with Claude Opus 5.5, it scores 3.0 to 7.6 points higher than Cursor, OpenCode, Devin, and Claude Code, at up to 24% lower cost. Droid 1.0 is available today with macOS and Windows desktop versions and a terminal UI, and also offers the Droid SDK (@factory/droid-sdk).

    Why read thisThe original article provides the specific mechanisms and measured data for programmatic tool calling and prompt caching, which can be used to understand where the cost and latency of long-task agents come from.

  12. Official Pydantic AI articleFirst-partyAnalysis
    AI score68

    Pydantic compares code mode and tool calling across 5,276 runs

    Michael Pfaffenberger of Pydantic used 5 small Pydantic AI applications and 8 models to conduct 5,276 scored runs, comparing code mode with direct tool calling. On long tasks, code mode reduced average cost by 58%, latency by 51%, and increased success rate by 16 percentage points; however, for short tasks with 5 items, 3 out of 5 tasks were instead 21% to 38% more expensive. The author argues that models are better at writing loops than at acting as the loop, and the gap widens as tasks get longer.

    Michael Pfaffenberger avatar

    Why read thisThe author compares code mode and tool calling across 5,276 runs, presenting curves showing how cost, latency, and accuracy vary with scale on long tasks.

  1. OpenRouter：AnnouncementsFirst-party
    AI score60

    OpenRouter will conduct CSAM screening on all inference requests

    OpenRouter announced that it will conduct child sexual abuse material (CSAM) screening on all inference requests during request processing; content flagged as suspected CSAM may be retained for manual review and reported and preserved in accordance with the law. Screening includes comparing content against hashes of known CSAM, and suspected content that does not match a hash but is flagged by classifiers will also be retained. This screening also applies to requests covered by OpenRouter's zero data retention (ZDR) policy; ZDR does not exempt the obligations to report and preserve in accordance with the law.

    Why read thisOpenRouter explained the mechanism for CSAM screening on all inference requests, as well as the boundaries under which ZDR traffic is also subject to it.

  2. Microsoft AI BlogFirst-partyModel releases
    AI score67

    Microsoft Releases Image Editing Model MAI Image 2.7, Ranked 3rd on Arena Edit

    Microsoft AI has released the image editing model MAI Image 2.7, launched in preview form, ranking 3rd on Arena Edit and 2nd on the Artificial Analysis image editing leaderboard, making Microsoft AI the 2nd-ranked lab. The company states that the model outperforms 2.6 on all 10 use cases and 7 categories of editing actions evaluated by Artificial Analysis, leading the next-best model by 24 Elo in scenes and styles. It is currently available for trial on Arena, Artificial Analysis, and MAI Playground, with text-to-image functionality coming soon.

    Blurry, light-colored flowers against a blue sky.

    Why read thisThe official rankings of this image editing model on Arena Edit and Artificial Analysis leaderboards are given, along with the improvement margin of use cases relative to 2.6.

  3. LangChain BlogFirst-partyAnalysis
    AI score67

    How LangChain Built the Design Agent Content Studio

    LangChain's GTM engineering team, working with the design team, built Content Studio, a design agent inside Slack, on Managed Deep Agents. It can generate videos, social graphics, UI mockups, and slide decks, and has already produced more than 3,000 assets and over 80 slide decks. The article breaks down its four-layer knowledge structure (skills, wiki, code validation, memory), the dedicated LangSmith Sandbox for each Slack thread, the two rendering engines HyperFrames and the in-house motion-studio, and how the main agent orchestrates asynchronous sub-agents. The team says that after migrating to Managed Deep Agents, its own code shrank by about 80%, the median model cost per request fell from about $8 in mid-September to about 50 cents, and the entire harness has been open-sourced.

    Shreya Talegaonkar

    Why read thisLangChain has publicly shared the complete build process for its design agent Content Studio, including the engineering trade-offs around skills, sandbox environments, and asynchronous sub-agents.

  4. Full Artificial Analysis articleFirst-partyAnalysis
    AI score82

    Gemini 4 Carbon Tops the Artificial Analysis Intelligence Index v5

    Google's Gemini 4 Carbon ranks first on the Artificial Analysis Intelligence Index v5 with a score of 59, leading in four of the 9 evaluations: HLE-Diamond, Terminal-Bench-Science, AA-Omniscience, and AutomationBench-AA. On the Coding Agent Index, Gemini 4 Carbon (high) paired with the Antigravity CLI scores 59, trailing Claude Code paired with Claude Opus 5.5 (max, 62 points) by 3 points. The model has a 1M-token context window, supports text, image, video, and speech input, and is priced at a standard rate of $4/$20 per million input/output tokens; it is currently available to select users with a 50% discount.

    Why read thisArtificial Analysis's independent evaluation places Gemini 4 Carbon at the top of the Intelligence Index v5 and provides a breakdown comparison with Claude Opus 5.5.

  5. OpenRouter：AnnouncementsFirst-partyProduct release
    AI score62

    OpenRouter Rebuilds Provider Routing Algorithm, Reducing TTFT by 16.9%

    OpenRouter announced that it has rebuilt its provider routing algorithm, shifting from allocating traffic mainly based on input token price to a real-time routing cost that comprehensively factors in actual price, time to first token (TTFT), throughput, error rate, and output quality. The company says the new algorithm reduces TTFT by 16.9%, increases throughput by 16.3%, and cuts 429 errors by 46%, while cost per request rises by 6.0%. Users do not need to change any code; explicitly setting provider.sort or provider.order can still bypass the new logic.

    How we rebuilt provider routing for better, faster inference

    Why read thisThe original article presents experimental data on switching between the old and new routing algorithms, so readers can use it to judge the trade-off between inference cost and latency.

  1. BaiduFirst-partyModel releases
    AI score62

    Baidu Open-Sources Unlimited OCR 2 Document Parsing Model

    Baidu has open-sourced Unlimited OCR 2, which uses a new SR-SWA mechanism to compress the attention interaction between text and vision into a fixed-size working reference, thereby parsing documents of almost any length within a limited context window. Baidu says that as documents grow longer, per-step decoding latency and memory usage remain nearly constant, and that the model achieves SOTA performance on MPDocBench and MarathonDocBench, which cover documents ranging from dozens to hundreds of pages.

    Why read thisThe official release presents the SR-SWA mechanism and results on two long-document benchmarks, offering insight into the context and latency trade-offs of long-document OCR.

  2. CursorFirst-partyProduct release
    AI score68

    Cursor Releases Agent SDK, Enabling Code-Defined Agents with Built-in Evals

    Cursor has released the Agent SDK, which lets you define agents in code, with prompts written in Markdown and tools written in TypeScript. Agents can be deployed on Cursor's managed infrastructure or run on your own, and built-in evals are used to measure each change and continuously improve the agent.

    Why read thisCursor packages agent definition, deployment, and evaluation into an SDK, allowing readers to gauge how the engineering cost of building their own agents will change.