New models, studies and security findings, read at the source
What a new model, paper or security result measured, how large the effect was and where the limits are.
- AI research and securityAnalysis1.9xMETR's estimate of how much faster Anthropic's benchmark scores rose
Is AI speeding up AI research? What METR found at Anthropic, and what it did not
METR's November 2026 review of Anthropic: benchmark scores growing 1.9 times faster, what that does and does not measure, its limits, and METR's time horizon.
- AI research and securityAnalysis89%Chinese share of open-source model spend, in Menlo's estimate
Open-source AI models in 2026: who leads, and what the 89% Chinese share means
Which open-weight AI models lead in November 2026, what Menlo's 89% Chinese share of open-source spend measures, and which releases are shipped or pending.
- AI research and securityGuide22.9%tool poisoning success in a 9,720-combination study
AI agent security in 2026: what the research shows goes wrong, and what to do
What 2026 research shows about AI agent security: tool poisoning, sabotage, exploit writing and the OWASP checklist, with the numbers, limits and steps to take.
The tools behind this research
AI Progress
FreeReleases, benchmarks and an outlook for the next 3 to 24 months.
See it liveAI Intel
FreeTen areas of AI, read in full and checked, one line per story.
See it liveAsk Spezia
Start hereFreeAn analyst that answers from verified data, with a source on every figure.
See it live