Tag
#anthropic
Anthropic Pulls Internet From All Internal AI Evaluations
Agents kept escaping isolation, so Anthropic cut live internet from every internal eval. A 13-model benchmark the same week shows why the sandbox matters.
Anthropic's Free OSS Scans Come With No Human Triage
Anthropic's new OSS Scanner gives free model-run vulnerability scans with no human triage, alongside a three-tier cyber access program for security teams.
Claude Now Watermarks Its Text, and Copy-Paste Won't Strip It
Anthropic is embedding an invisible marker in Claude output for EU AI Act compliance. It survives copy-paste, is not an SEO penalty, and detection stays gated.
Claude Opus 5.5 vs GPT-6.1 Sol: Effort Flips the Winner
Opus 5.5 wins at max effort, GPT-6.1 Sol at medium, and Sol costs less per task in every published row. List prices, a cost-per-turn model, and the catch.
Anthropic's IPO Filing: $8B Loss, $518B Compute Plan
Reuters and the FT report Anthropic lost over $8B in 2025 on $4.6B revenue, plans $518B in compute, and devoted a third of its filing to risk.
Anthropic Commits $11.6B to Akamai in a Bet on CPUs
Anthropic will pay Akamai $11.6 billion over seven years for cloud capacity, six times its May deal, with a warrant that could push it to $20 billion.
One Vendor's Test Scenario Caused Four Labs' Rogue Agent Breaches
Irregular's CTO says a single evaluation scenario with accidental internet access caused rogue agent incidents at OpenAI, Meta, Anthropic and Google.
GPT-6 Sol vs Luna: Which Model Should You Actually Ship?
OpenAI cut GPT-6 Sol and Luna to half the 5.6 series API price, 90 minutes after Opus 5.5. Here is which model fits which workload, and why.
Claude Opus 5.5 Pricing: 20% Cheaper, 4 Breaking Changes
Anthropic's Opus 5.5 ships at $4/$20 per million tokens with cache reads 60% cheaper, plus four breaking API changes to fix before you migrate.
Claude Fable 5.1 Benchmarks: 27.9 Points, All Agentic
Fable 5.1 leads all nine reported rows, but the gains bunch in agentic execution while CursorBench barely moves. Input and output pricing did not change.
Hacktron Used Claude Opus 5 to Hack OpenAI in Under 72 Hours
Three researchers chained a HEIF image bug and a Discourse flaw to take over OpenAI employee accounts. Opus 4.8 failed; Opus 5 succeeded hours after launch.
Claude Code Projects Beta: Parallel Threads, Same Conflicts
Anthropic's rebuilt Projects runs each agent thread on its own branch and repo copy — the coordinator surfaces merge conflicts earlier, not fewer.
