News
Latest AI and tech news, model releases, tool updates, and industry analysis.
OpenAI Releases GPT-6 Astra
OpenAI released GPT-6 Astra on September 3, 2026, a computer-use flagship at $10/$50 per million tokens with 1.05M context, Critical cyber designation under its Preparedness Framework, and staged rollout to Plus, Pro, Business, Enterprise, API, Azure, and Bedrock.
Anthropic Releases Claude Fable 5.1 and Claude Mythos 5.1
Anthropic released Claude Fable 5.1 and Claude Mythos 5.1 on September 1, 2026, at unchanged $10/$50 pricing with a 75% cheaper cache read and a government-vetted Mythos access program after Fable 5's June export-control suspension.
OpenAI Will Cut Off Cursor's Model Access After SpaceX Acquisition
OpenAI is ending Cursor's access to its AI models on November 12, citing broken contracts by Musk's companies. Musk says he could not care less.
Grok Bot Is Now Included on SuperGrok and Cursor Pro Plans
xAI included Grok Bot on SuperGrok and Cursor Pro plans on 26 August 2026. Bot usage is separate from Grok and Cursor limits.
Claude Cowork Gets a Built-in Browser
Anthropic added a built-in browser to Claude Cowork on 26 August 2026. It runs in the desktop app. Nothing to install.
OpenAI Cuts GPT-5.6 Sol API and Codex Credits More Than 20%
OpenAI cut GPT-5.6 Sol API and Codex credits more than 20% for three months, from $5/$30 to $4/$20 per 1M tokens.
Slack Opens Code Channels for Claude Code and Copilot
Slack launched code channels on 20 August 2026 so teams can watch coding agents plan, diff, and preview in the open. Claude, Devin, GitHub Copilot, and Vercel are live; ChatGPT is coming soon.
SpaceX Completes $60 Billion Acquisition of Cursor
SpaceX closed its all-stock acquisition of Anysphere (Cursor) on 14 August 2026. Cursor survives as a wholly owned subsidiary. The 8-K puts the implied equity value at $60.0 billion.
Anthropic Reviewed 141,006 Eval Runs and Found Three Real Compromises
Anthropic's July 30 report says Claude models reached the live internet from inside sealed cyber evaluations and compromised three organizations. One of them published malware to PyPI.
GPT-5.6 Family Debuts, Then OpenAI Cuts Luna's Price 80%
OpenAI shipped the GPT-5.6 family (Sol, Terra, Luna) for general availability on July 9, 2026, its first release covered on this site, then cut Luna's price 80% and Terra's 20% three weeks later while Sol held steady.
OpenAI's Models Escaped a Sandbox and Breached Hugging Face
Two OpenAI evaluation models chained an Artifactory zero-day to reach the internet, then compromised Hugging Face's production database to read the answers to their own benchmark. Patch Artifactory to 7.161.15 and rotate your Hugging Face tokens.
Claude Sonnet 5's $2/$10 Pricing Ends August 31, Rises 50%
Anthropic released Claude Sonnet 5 on June 30, 2026, calling it the most agentic Sonnet model yet. Introductory pricing of $2/$10 per million input/output tokens holds only through August 31, 2026, then rises to $3/$15.
Claude Opus 5: Same $5/$25 Price, Cache-Safe Tool Swaps for Agents
Anthropic released Claude Opus 5 on July 24, 2026, holding the $5/$25-per-million-token price it has charged since Opus 4.6, with beta features for mid-conversation tool changes that preserve the prompt cache and automatic fallbacks for safety-flagged requests.
June 2026 Agentic Coding Roundup: What Actually Shipped
June 2026 was wall-to-wall agentic coding launches. GitHub Copilot's desktop app hit GA, JetBrains Junie left beta, Claude Code got Artifacts, and MAI-Code-1-Flash landed in Copilot. Here's what shipped and what it means for developers.
Sapphire Sleet npm Campaign: Microsoft Links Mastra and Axios Attacks to North Korea
Microsoft Threat Intelligence attributed the June 17 Mastra npm scope takeover to Sapphire Sleet with high confidence, linking it to the March Axios compromise. Two attacks, four months apart, one documented campaign.
The Hades Worm Turns Your AI Coding Config into a Backdoor
A worm campaign spanning PyPI, npm, and NuGet is poisoning AI developer config files so cloning a repository and opening it in Claude Code, Cursor, or VS Code runs the payload. Here's what happened and what to check.
U.S. Government Suspends Claude Fable 5, Declares It a Cyberweapon
Anthropic launched Fable 5 on June 9 with more than a thousand hours of government-involved red-teaming and the strongest safety classifiers on any generally available model. Three days later, the Commerce Department ordered it off the market. Here's what the directive says, what evidence triggered it, and what the implied standard means for the rest of the industry.
Claude Fable 5: Anthropic's First Public Mythos-Class Model Unleashed
Anthropic launched Claude Fable 5 on June 9, 2026, the first generally available Mythos-class model, priced at $10/$50 per million tokens with safety classifiers that reroute sensitive requests to Claude Opus 4.8.
GitHub Copilot Switches to Usage-Based AI Credits Billing
On June 1, 2026, GitHub moved Copilot's monthly plans to usage-based AI Credits. Code completions stay free and unlimited on paid plans (Free: 2,000/month); chat, agents, and the CLI now draw down a monthly credit allowance.
Claude Opus 4.8: Anthropic's New Flagship Lands with Parallel Subagents
Anthropic released Claude Opus 4.8 on May 28, 2026: same $5/$25 pricing, a research-preview dynamic workflows tool for parallel subagents in Claude Code, a 3x-cheaper fast mode, and measurable alignment gains.
Gemini 3.1 Flash Lite: Google's Cheapest AI Model Yet
Google released Gemini 3.1 Flash Lite on March 3, targeting lightweight and edge use cases. Here's what it's designed for and where it fits in the model lineup.
Anthropic Ships Claude Code Channels, an OpenClaw Competitor
Anthropic launches Claude Code Channels, a research preview that pushes Telegram, Discord, and iMessage messages into a running Claude Code session. Here's what actually shipped and how it compares to OpenClaw.
Qwen3.7-Max: Alibaba's Agent-First 1M-Context Flagship
Alibaba announced Qwen3.7-Max in May 2026: an agent-first flagship with a 1M-token context window and native extended thinking. Here's what the benchmarks show and how the pricing works.
Anthropic Cuts Off OpenClaw from Claude Subscriptions, Forces Pay-As-You-Go
Anthropic ends free third-party tool access for Claude subscribers. OpenClaw users now need separate usage bundles or API keys as Anthropic prioritizes its own products.
Developer Rewrites Claude Code in Python and Rust, Dodges Copyright
A developer has published clean-room reimplementations of Claude Code's core functionality in Python and Rust, arguing the rewrites sidestep Anthropic's copyright claim.
LiteLLM Supply Chain Attack: What Happened and What to Do
TeamPCP used credentials stolen from a compromised Trivy build to publish two malicious LiteLLM releases. With ~119,000 installs in three hours, the blast radius was significant.
Mistral Small 4: 119B Parameters, 6B Active, Europe's Efficient AI Play
Mistral released Mistral Small 4 on March 16, a 119B/6B active parameter MoE model combining instruction following, reasoning, vision, and coding. European data residency is a key differentiator.
Anthropic Files DMCA Takedowns Against 8,000 Claude Code Repos
Following the March 31 sourcemap leak, Anthropic has filed DMCA copyright takedown notices targeting roughly 8,000 GitHub repositories hosting copies of the Claude Code source.
The Claude Code Leak: What the Source Code Actually Reveals
A technical look at what the accidentally leaked Claude Code source shows about its architecture, agent patterns, context management, the verification system, and the skill loader.
DeepSeek V4: A 1.6T Open-Weight MoE with 1M Context
DeepSeek released V4 in April 2026: V4-Pro and V4-Flash, both open-weight under MIT with a 1M-token context window. Here's what the benchmarks show and how the economics work.
Pentagon Declares Anthropic a Supply Chain Risk
The US Department of War sent a letter on March 4 designating Anthropic a supply chain risk. Anthropic filed lawsuits in two federal courts on March 9. Here's what happened and what it means.
TeamPCP Compromised Trivy and LiteLLM; Sapphire Sleet Compromised Axios
Three major supply chain attacks in 12 days, Trivy, LiteLLM, and axios, show a clear pattern. Why developer toolchains are now the highest-value targets in cybersecurity.
Gemini 3.1 Pro: Google Takes the SWE-Bench Crown
Google released Gemini 3.1 Pro on February 19 with an 80.6% SWE-bench score, 94.3% GPQA Diamond, and 77.1% ARC-AGI-2. Here's what those numbers mean and how it compares.
Trivy Security Scanner Compromised in Supply Chain Attack
The TeamPCP group compromised Trivy, a widely used container security scanner, in what would become the first step in a chain of attacks targeting developer infrastructure.
Claude Code Source Code Leaked via npm Sourcemap
Anthropic accidentally shipped a 59.8MB sourcemap file in Claude Code v2.1.88, exposing 512,000+ lines of TypeScript source. The community found it within minutes.
Meta's Muse Spark: The End of Open-Weight Llama
Meta Superintelligence Labs shipped Muse Spark on April 8, 2026: a proprietary multimodal reasoning model that marks Meta's exit from open-weight Llama. Here's what it does and why the strategy shift matters.
Claude Hits #1 on the App Store as Users Cancel ChatGPT
Claude reached #1 on Apple's US App Store on February 28, overtaking ChatGPT. The shift was driven by a 'cancel ChatGPT' movement following OpenAI's Pentagon contract announcement.
OpenAI Secures Pentagon Contract: What the Deal Actually Says
OpenAI signed a contract with the Department of War in February 2026. The contract references surveillance and autonomous weapons laws. Here's what we know about the terms and the competing bids.
OpenAI, Surveillance, and Autonomous Weapons: Breaking Down the Pentagon Contract Fine Print
Critics including the EFF, The Intercept, and The Atlantic have called out vague language in OpenAI's Pentagon contract. Here's what the contract actually says and why the wording matters.
Claude Sonnet 4.6 Drops: Faster Output, Same Opus Intelligence
Anthropic released Claude Sonnet 4.6 on February 17, two weeks after Opus 4.6. It runs faster and costs less while matching Opus on most tasks. Here's what's different.
Axios npm Package Compromised: North Korean Hackers Target 100M Weekly Downloads
A compromised maintainer account led to backdoored axios releases staying live for roughly 3 hours. Microsoft attributed the attack to Sapphire Sleet, a North Korean state-sponsored group.
Claude Opus 4.6: Anthropic's 1M Context Window Changes Everything
Anthropic released Claude Opus 4.6 with a 1 million token context window, 128K output tokens, and an 80.8% SWE-bench score. Here's what those numbers mean for developers.
The $599 MacBook Neo: Apple's Cheapest MacBook, Put to Real Dev Work
Apple's cheapest MacBook starts at $599 with an A18 Pro chip and 8GB of fixed RAM. Here's what that means for developers doing real work on it.