A research team from alphaXiv, Stanford, MIT and other institutions ran a 12-week experiment placing 936 agents in a shared preprint server and chat space, with 35 human labs supplying only compute, API budgets and broad research directions. The agents autonomously generated hypotheses, wrote and ran experiments, analyzed data and authored papers, producing 936 papers and 9,500 messages across 229 self-created chat rooms, while spontaneously developing replication checks, peer review and even their own journal. The platform is now open for external researchers to connect their own agents.
AI for Science
Applying AI across the scientific discovery pipeline, from literature review and hypothesis generation to experiment design and
Curated Intelligence
Alibaba unveiled Qwen4-Max, a 2.4T-parameter base model with 71B activated parameters built on a new QSA-based MoE architecture, featuring native vision and million-token context. It cuts average cost per task by 58% versus its predecessor, ranks sixth globally in the CodeArena front-end blind test, and is now available via API on the Qwen AI platform, with smaller variants to be open-sourced.
Moonshot AI is redirecting its Kimi models from coding toward biology, chemistry, and pharma research, citing diminishing returns in coding benchmarks. International director Giacomo Ficari said at Web Summit Lisbon that science demands smarter models than law, finance, or code. DP Technology has integrated Kimi K3.1 Pro into its BioMaster research assistant, where one test analyzed gene data from 3,696 mouse pancreatic cells and flagged 100 candidate genes in 90 minutes versus an estimated four hours of human work, though fully independent breakthroughs remain distant.
OpenAI is scaling its ChatGPT for Academic Researchers program to up to 200,000 researchers worldwide, with the next application round opening November 23 and 50,000 new grants awarded first-come, first-served, each covering the recipient plus up to two collaborators on ChatGPT Plus. About 20,000 researchers across 33 countries have joined since the July pilot, and OpenAI says program data is not used to train its models by default.
OpenAI's Economic Research team analyzed over 580,000 messages from 1,819 scientists in its ChatGPT for Academic Researchers summer pilot and surveyed 251 participants. Scientists used ChatGPT across literature review, data analysis, coding, and idea generation; respondents estimated the same-quality work would take 1.9x longer without AI, and 98% expected AI to accelerate discovery in their field.
Eli Lilly CEO Dave Ricks publicly rebutted Anthropic CEO Dario Amodei's prediction that AI will cure cancer within a decade, calling the vision appealing but unlikely. He argued that major gaps remain in biological understanding and that randomized human trials, not drug discovery speed, are the real bottleneck, so AI will likely deliver steady incremental progress rather than a sudden breakthrough.
Researchers used AI to transcribe and formalize 1,009 pages of Kurt Gödel's densely written notebooks, much of it in an obsolete shorthand. The archive is said to contain hundreds of previously unpublished proved theorems, marking a landmark intersection of mathematical history and AI for Science.
Claude Fable 5.2 independently completed a tedious black hole formation calculation that physicists had left unfinished since 2003, finding that black holes form suddenly in tiny universes just as Hawking and Page showed for large ones in 1983. A physicist has verified the work, marking a case where an AI produced checkable new knowledge in fundamental physics rather than merely summarizing existing text.
AI-for-science startup Physical Superintelligence used an internal research system running a non-public Anthropic model to find a complete proof of the long-standing Aaronson-Ambainis quantum speedup conjecture within three days, validated via the Lean proof assistant. It marks a landmark case of AI autonomously extending human knowledge, foreshadowing synthetic researchers outpacing human comprehension.
Anthropic is inviting principal investigators at academic institutions, nonprofits, and government labs to apply for an 8-12 week guest investigator program, working directly with Anthropic researchers to apply Claude to causal human biology, structural and computational molecular biology, medicinal chemistry, drug discovery, and clinical science. Anthropic will run its models and harnesses at no cost, with outputs intended for joint publication.