Replit Agent's new harness design beats the sidekick architecture by 11 points on DeepSWE and 16 on Terminal-Bench, achieving Pareto-efficiency against Astra baselines.
Anthropic reports that Zhipu AI's GLM-5.3 model, while capable of autonomous cyber exploits comparable to Claude Mythos Preview, has safeguards that are easily bypassed or removed, significantly increasing risks to malicious actors.
Kumo Tabular achieves state-of-the-art accuracy on TabArena with ELO 1950 while being 17x faster than LimiX-2, trained entirely on synthetic data.
GitHub's H1 2026 transparency report shows a rise in government takedown requests to 708, attributed to expanded reporting methodology rather than increased moderation.
Quine combines a multimodal world model with an interactive harness to prioritize therapeutic compounds, validated in collaboration with the Broad Institute.
Researchers developed a method to estimate and rank the energy consumption of proprietary AI models from Anthropic, OpenAI, and Google, which do not publicly disclose this data.
The new verification layer caught 138 of 139 expert-flagged unsupported claims in medical agent tests while identifying the correct source for 86% of claims with identifiable sources.
The new module supports PCIe cascading and runs on embedded Linux and Android systems.
Cloudflare developed CryptoLabe, an internal AI tool built on its Developer Platform, to scan codebases and chart a course for its 2029 post-quantum readiness deadline.
Reclaim uses a custom agent loop and 'Preview Mode' to let users review AI-suggested calendar changes before they affect other attendees.
Cloudflare has implemented a beta mitigation for an IPsec design flaw enabling quantum downgrade attacks, with support rolled out for Cloudflare WAN and Magic Transit.
Cloudflare has launched Application Profiles, a new feature that extends Schema Learning and Validation to web applications, allowing customers to enforce positive security policies by blocking traffic that deviates from learned request structures.
Customers can now inspect per-connection post-quantum adoption via Logpush, Log Explorer, and HTTP Traffic Analytics to audit compliance and identify cryptographic gaps.
Cloudflare has launched Threat Signals, a new agentic tool that automates the conversion of open-source threat intelligence into actionable WAF rules and indicators, now available for free to every Cloudflare account.
The test generated 1,107 mutation attempts, resulting in 49 validated findings that led to new SSRF detection rules in the Managed Rules update.
Cloudflare will offer standard Merkle Tree Certificate issuance at no cost, targeting inclusion in Chrome’s Quantum-resistant Root Store by early 2027.
Cloudflare introduces an adaptive application security framework featuring LLM-hardened WAF rules, free threat intelligence access, and new tools for governing agentic traffic.
Cloudflare has applied for inclusion in major root programs and signed an agreement to acquire a GlobalSign root to ensure broad device compatibility.
ElevenLabs has joined the newly launched OpenAI Marketplace as a launch partner, allowing eligible enterprise customers to access ElevenAgents and apply their OpenAI commitments.
Jev achieves 96.47% accuracy on IMDb with a cost of $0.6492, raising the bar for justifying custom fine-tuning.
A Markdown-based contract language that declares states and conditions instead of execution order, re-running only upon changes to reduce costs.
The model features 320B total parameters with only 15B activated per token, optimized for reasoning and multi-step tool use.
OpenAI opened ChatGPT as a shared surface for developers to launch native experiences to 1.2 billion weekly users.
AI Briefing