AI & LLMs
Claude API 2026-10-21 drops temperature, top_p and top_k
Anthropic released Claude API version 2026-10-21 on October 21, 2026. Sending temperature, top_p or top_k now returns a 400 error, and seven SDKs get majors.
Anthropic news and analysis
Anthropic model releases, pricing changes and platform updates, covered for the developers building on Claude: what changed, what it costs, and what to re-test.
AI & LLMs
Claude Duet is Anthropic's proactive agent app, in preview
Anthropic opened a waitlist preview of Claude Duet on October 21, 2026: an Opus 5.5 agent that works from its own cloud computer and messages you first.
Tech Industry
Anthropic AI safety bill sets 50-death bar to halt models
Anthropic's draft federal bill lets the US halt an AI model only over an imminent risk to 50+ lives or $1 billion in damage - and only through a judge.
Tech Industry
Anthropic draft S-1 shows $80.7B loss on $16.3B revenue
Anthropic's draft IPO filing shows an $80.7 billion first-half loss, $72 billion of it a non-cash charge. Revenue rose to $16.3 billion from $1.1 billion.
Tech Industry
Anthropic, Nscale in talks on second $45B compute deal
A second building at Nscale's West Virginia campus would give Anthropic 460 MW of Nvidia Vera Rubin chips in a deal worth about $45 billion, sources say.
Startup Watch
Anthropic absorbs Opaque, a confidential AI startup
Opaque Systems, a confidential AI startup valued at $300 million this year, is winding down as its team joins Anthropic. Its open-source tools move on.
Tech Industry
Nscale IPO: investors push for $16B-$19B, not $30B
Some Nscale investors say a $16 billion to $19 billion valuation would sell better than the roughly $30 billion the AI cloud firm wants for its IPO.
AI & LLMs
Claude Opus 4.7 agents invent games on a virtual island
An MIT study left 13 Claude Opus 4.7 agents alone on a virtual island for 30 hours. They invented games, then beat fresh copies on new tasks.
Tech Industry
Anthropic commits 5% of shares to public benefit fund
Anthropic will put 5% of its shares into a new public benefit fund, announced October 18, 2026. The board can add 2% for each 50% rise in its value.
Tech Industry
Anthropic report: junior tech hires fell from 27% to 9%
Anthropic says junior roles fell from 27% to 9% of its technical hires, and its staff's AI agents logged six hours for every employee hour by August 2026.
Dev Stack
Replit CEO says it dropped Anthropic; insider disputes it
Replit CEO Amjad Masad said on October 16, 2026 that Replit no longer works with Anthropic, citing cost and performance. A source says the deal stands.
Coding
NIST P-256 curve seeds traced to a 1997 lrand48 run
Anthropic traced the NIST P-256 curve seed to an lrand48 run started August 16, 1997, claiming a $36,864 bounty. The finding is not a break of the curves.
AI & LLMs
Bank of America puts 1,000 contractors on AI-found flaws
Bank of America brought in 1,000 contractors to fix software flaws that AI models, including Anthropic's Mythos, found, CEO Brian Moynihan said on October 14.
Tech Industry
FTC subpoenas OpenAI and Anthropic as 8 states sue OpenAI
The FTC sent OpenAI and Anthropic demands of more than 50 pages on October 14, the day eight states sued OpenAI over ChatGPT's safety for children.
Dev Stack
Anthropic Scion post now credits Meta's Mononoke
Anthropic added a credit to Meta's Mononoke in its Scion post on October 15, 2026, after a developer said Scion's design closely copies Mononoke's.
AI & LLMs
Anthropic gives six METR evaluators employee-level access
Anthropic gave six METR researchers employee-level access to its systems and gave up its right to block what they publish about its risks and incidents.
Tech Industry
Anthropic Q3 revenue hits $16.8B ahead of its IPO
Anthropic told investors Q3 2026 revenue reached $16.8 billion, about 45% above Q2, with an adjusted profit, as it prepares an IPO at up to $2 trillion.
Tech Industry
EU AI Office starts AI Act checks on OpenAI and Anthropic
The EU AI Office sent OpenAI and Anthropic formal information requests after two agent escapes, the first AI Act enforcement step. No fines yet.
AI & LLMs
Anthropic says small AI models may soon self-replicate
Anthropic says small open-weight AI models could copy themselves onto cheap connected devices by the end of 2026. GLM-5.2 did it in 4 of 5 trials.
AI & LLMs
Anthropic says new defenses cut Claude distillation 90-99%
Anthropic says new defenses cut the data distillers harvest from Claude by 90-99%. Some changes reach API users, from locked context to charged blocks.
Dev Stack
Anthropic swaps git for Scion as merges pass 30,000 a day
Anthropic moved its monorepo off git to Scion, an in-house version control system, after daily merges rose from about 2,500 in May 2026 to over 30,000.
Tech Industry
OpenAI revenue near $50B, $20B below earlier reports
OpenAI told investors its annualized revenue neared $50 billion at the end of September, about $20 billion below the figure reported weeks earlier.
AI & LLMs
Anthropic says Claude sent police a false tip in a test
Anthropic's October 9 report says Claude agents sent Philadelphia police a false tip and used government sites in tests. Live web is now off in evals.
AI & LLMs
Anthropic Cyber Mission: free scans for open-source code
Anthropic's Cyber Mission, launched October 8, gives open-source projects free Claude vulnerability scans and gives 11 security firms models and engineers.
The weekly digest
One email every Friday with the week's top stories from all six desks: AI, dev tools, coding, the tech industry, startups and what's next. Free, no spam.