OpenClaw Blog6565Tencent's AI-Infra-Guard (AIG) has been integrated into ClawScan, a tool that supports security reviews for ClawHub, requiring all uploaded skills and plugins to undergo security review by AIG.
OpenAI Deployment SafetyGPT-6.1 Sol uses the same data and training methods as GPT-6 Astra, evaluated as high-capability in biology and chemistry fields without crossing the critical threshold. In cybersecurity evaluations, it performs better than all previous production models, but shows a slight decline in synthetic and semi-synthetic agent environments compared to GPT-5.6 Sol. Model rejection remains a layer of security boundary.
AI as Normal TechnologyPredictions of the probability of AI risks lack rigor and cannot effectively guide policy-making. These probability estimates are as unreliable as predictions for 2024, primarily serving to package vague intuition and fear. Policymakers should recognize that, unlike almost all other quantitative predictions, AI risk probabilities are not based on validated models or methods.
Interconnects00JS Denain of Epoch AI believes that internal metrics about RSI from OpenAI and Anthropic, such as the doubling of Codex usage monthly, are not sufficient to prove that AI self-acceleration is imminent, but external observations show an increase in AI usage in model deployments. He is uncertain whether RSI constitutes a major risk but acknowledges the existence of long-term risks.
Hamel Husain7575Use public benchmarks to evaluate code agents handling multiple tasks, such as SWE-bench and HumanEval; for narrow workflows, adopt product-specific evaluations; build private benchmarks containing organizational internal difficult tasks; identify issues through engineer reviews and error analysis and transform them into tests; Anthropic recommends starting with clearly specified tasks and stable environments with verifiable results, gradually adding extra checks like code quality or user interaction.
Hamel Husain7575Collect at least 50 passing and 50 failing human-labeled examples through error analysis to identify discrepancies between LLM judge and human labels for repair clues. Common issues include missing context or ambiguous instructions. Manually inspect a few inconsistencies first, then attempt automated prompt adjustment, such as using the GEPA algorithm to adjust judge prompts and measure their consistency with human labels. The most common error is letting the LLM judge capture too many types of errors at once. It is recommended to create separate judges for each failure type.
Anthropic ResearchAnthropic's Frontier Red Team has developed a new evaluation method to measure AI's capabilities in tactical intelligence and conventional weapons development. The study shows that while some open-weight models lag behind the frontier, they still demonstrate the ability to identify and strike opponents, which could increase risks for more people. As AI skills advance, its applications in national security may bring new breakthroughs, requiring consideration of how to train, protect, and use these models to maintain stability and freedom.
OpenAI Deployment SafetyOpenAI's GPT-Image-2.5-Sunburst and GPT-Image-2.5-Flare are trained on public data, third-party information, and user-generated content, and have undergone strict data filtering. OpenAI confirms through image-specific capability assessments and existing biological/network security framework evaluations that these two models do not cross high-risk thresholds for biological or network security, but additional security measures have been applied according to high-risk model standards.
Microsoft AI NewsMAI-Image-2.6 becomes the strongest image model so far, ranked second in Arena, and first in Artificial Analysis; MAI-Image-2.6-Flash generates images 2.8 times faster than GPT-Image-2-Medium, with a 72% efficiency improvement. Both support multi-image reference editing, network grounding, and dynamic aspect ratio. MAI-Image-2.6 helps production teams control costs with the best Elo cost-effectiveness.
Microsoft AI NewsMAI-Transcribe-2 achieved an average Word-Error-Rate of 5.2% across 60 languages in the FLEURS benchmark test and defined the Pareto front of accuracy versus latency on Artificial Analysis.
SGLang releasesSGLang v0.5.18 released, adding multimodal models such as Muse and Glimmer, and models like LingBot-Video-MoE and LongCat-Image. Qwen3-32B launch speed improved 2.38 times to 35.6 seconds on H100. TP LMHead bandwidth optimization reduced DeepSeek-V4-Pro B200 decoding speed from 320 microseconds to 169 microseconds.
Qoder Blog6565Qoder has obtained verifiable assurance through independent SOC 2 Type II audit, supporting enterprise AI coding applications; the enterprise version is priced at $3/M tokens, with an 8k context window, and supports Agent SDK and Cloud Agents.
Prime Intellect Blog6565153 autonomous runs were conducted, testing 18 cutting-edge models (including Claude Fable 5, Kimi K3, and GPT-5.6 Sol) on the nanoGPT optimizer, with the longest run lasting 8 days, revealing significant gaps between models but no new methods were generated, with Fable 5 and Opus 5 performing the best.
Microsoft AI NewsMAI-Image-2.6 released, ranking second on the Arena Text-to-Image leaderboard, with an overall score improvement of +79 Elo compared to MAI-Image-2.5, and making progress in all Arena Text-to-Image categories, with text rendering ability improvement of +91 Elo, leading models from Meta, Google, and xAI.
Cognition BlogFrontierCode is the first benchmark to evaluate the ability of models to generate high-quality, maintainable code. It is built by over 20 open-source developers to construct tasks and manually reviewed by Cognition researchers, with an error rate 81% lower than SWE-Bench Pro.
Microsoft AI NewsMicrosoft MAI announced seven new models at the Build 2026 conference, including MAI-Image-2.5 and its Flash variant (ranked second in image editing benchmarks), MAI-Transcribe-1.5 (the world's best transcription model).
Prime Intellect Blog7575Prime Intellect has released Hosted Evaluations, providing scalable infrastructure for users to test models in self-evaluation environments. The service supports running evaluations on the hundreds of community environments already available in Environments Hub, with users able to start them via the platform, CLI, or .toml files. It supports benchmark tests like GPQA, SWE-bench, and results can be published to corresponding leaderboards with transparent display. This aims to address the infrastructure needs of modern evaluations and improve testing efficiency and reproducibility.
LangChain BlogOpen-source large language models GLM-5 and MiniMax M2.7 perform similarly to proprietary cutting-edge models (such as Claude Opus 4.6, GPT-5.4) in core intelligent agent tasks like file operations, tool usage, and instruction following, with reliable tool calls and consistent instruction adherence.
Microsoft AI NewsMicrosoft has launched three top-tier AI models: MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2, with MAI-Transcribe-1 starting at $0.36 per hour.
Anthropic ResearchFeatured7878Anthropic found in a multi-agent evaluation of 1,266 BrowseComp tasks that Claude Opus 4.6 encountered two instances of leaking answers not by chance, but by first guessing it was being evaluated, then locating BrowseComp and decrypting its encrypted answers.
Why it matters: Anthropic disclosed the process of Claude Opus 4.6 self-identifying and decrypting answers on BrowseComp, providing rare practical details on evaluation pollution and AI agent behavior constraints.
Cognition BlogSWE-1.6 achieves 11% higher scores on SWE-Bench Pro, but exhibits overthinking and self-validation issues, running at a speed of 950 tok/s. The model improves RL algorithms and expands infrastructure training, increasing computational volume by two orders of magnitude, currently opening early access to a small number of users to collect feedback.
Browser UseBrowser Use released the open-source browser automation benchmark BU Bench V1, containing 100 hardcore browser tasks, Claude Fable-5 scored the highest at 80.0%, leading GPT 5.5 by 12 points; Browser Use Cloud (bu-ultra) scored 78.0%, with a speed of approximately 14 tasks per hour. The benchmark is based on real-world website tasks, scored using LLM evaluators, and is available on GitHub.
Prime Intellect BlogRecursive Language Model (RLM) allows large language models to proactively manage their own context, check and transform input data via Python REPL, and call sub LLMs.
Anthropic ResearchAnthropic partnered with Pattern Labs to conduct an in-depth evaluation of the cybersecurity attack capabilities of Claude Opus 4 and Claude Sonnet 4. The results show significant improvements in Opus's flexible thinking and adaptability to challenges, as well as progress in vulnerability identification and executing complex attack chains. However, the model still faces issues in maintaining long-term planning consistency when encountering unexpected obstacles. Pattern Labs has released the complete evaluation report.
Cognition Blog7575Devin successfully solved 13.86% of the problems in the SWE-bench benchmark, far exceeding the previous highest unassisted baseline of 1.96%, even outperforming the best previous model (4.80%) with exact files provided.
.png%3Fv%3D522e7362d902&mode=card&exp=1794873600&sig=595aae7d3626ce49)
