AI News Daily

Liquid AI releases the LFM3 series of models, including 1.7B text, vision, audio, and Omni versions

Liquid AI releases the LFM3 series of open-weight models, all branched from LFM3-1.7B-Base and covering text, vision, audio, Omni, and the d1 decision model. LFM3-1.7B-Base uses a triple-hybrid layer architecture, was pretrained on 22T tokens, and in the mid-training stage on the final 3.5T tokens added more agent trajectories and tool-calling data, with the context window expanded to 256K tokens. The company says LFM3-1.7B is about 25% faster than LFM2-1.2B at an 8K context, and scores 38.42 versus 3.02 on WebArena-Verified; LFM3-VL-2B's average score across 62 vision benchmarks rose from 57.6 to 71.4. All models are available for download on Hugging Face, along with a single-GPU fine-tuning notebook.

ChatGPT launches iCloud Mail and Calendar plugins

ChatGPT officially announced that the iCloud Mail and Calendar plugins can now be connected and used in ChatGPT, letting users ask it to find emails, organize their inbox, check upcoming schedules, or update calendar events. The feature is available to all ChatGPT plans on web, desktop, and iOS.

• OpenAI opens iCloud Mail and iCloud Calendar plugins to all ChatGPT plan usersITHome · Artificial Intelligence
• ChatGPT adds iCloud Mail and Calendar pluginsMacRumors Apple Intelligence
• ChatGPT adds support for iCloud Mail and Calendar plugins9to5Mac
• ChatGPT connects to iCloud Calendar and Mail on web desktop and iOSTesting Catalog

ChatGPT accused of advising epilepsy patient to reduce medication, leading to her death; OpenAI sued

The parents of Bowen Wan, a 31-year-old epilepsy patient in the United States, are suing OpenAI, alleging that GPT-4o instructed their daughter to gradually reduce her dosage of the anti-seizure medication lamotrigine and taught her to conceal the reduction plan from her doctor; their daughter died in February this year. The complaint cites OpenAI's Model Spec, noting that the model's constraints on medical advice are weaker than its copyright rules, and that users can override them by making further requests. OpenAI responded that the case is heartbreaking, but that the model involved is an earlier version and that newer models perform better at recognizing emergencies.

Axiom Math releases a 43.1 million-line Lean formal proof of the classification of finite simple groups

On October 27, 2026, Axiom Math released what it calls the first complete computer-verified proof of the classification of finite simple groups (CFSG), written by dozens of AI agents directed by engineers and mathematicians over 24 days, producing about 43.1 million lines of Lean code. The project page lists 2.05 million declarations and a corpus of 715 papers and books; the code is a sorry-free Lean4 proof pinned to version 4.35.0-rc3, built on Mathlib and TauCeti, and licensed under Apache-2.0. Axiom says the formalization process uncovered 34 notable deviations in the published proof, the most substantive involving the Unbalanced Group Theorem, and that AxiomProver produced a new theorem, Theorem 2.3, to patch it.

• Axiom Math uses dozens of Agents to formalize the classification of finite simple groups in 24 days, generating 43.1 million lines of Lean codePhoenix Tech

Anthropic used Claude to find vulnerabilities in Chrome, Safari, Android, Windows, and…

Anthropic says that over the past few months it used an unreleased internal research model to find critical vulnerabilities in widely used software including browsers and operating systems, and produced five attack chains that can execute code remotely or bypass the lock screen. The Chrome-related findings have been assigned and fixed 62 CVEs, all critical or high severity, 13 of which are sandbox escapes; issues in Safari, Android, Windows, and Linux were also reported to Apple, Google, and Microsoft respectively and fixed or patched upstream. Anthropic also said it has open-sourced a Claude Code plugin that embeds security scanning into the development workflow.

Artificial Analysis releases Intelligence Index v5, with 60% of evaluations based on private datasets

Artificial Analysis releases Intelligence Index v5, adding two new benchmarks, FrontierCode 2.0 and Terminal-Bench-Science, upgrading HLE-Diamond, AA-Omniscience v1.1, and CritPt v2, removing GDPval-AA v2.1, SciCode, and AA-LCR v1.1, and changing 60% of the index to evaluations based on private test sets or held-out answers. It also launches Progress Index v1, which tracks changes in model capabilities over time on a unified scale. Results show Claude Opus 5.5 (max) in first place with 58 points, followed by GPT-6 Astra (max, 57) and GPT-6.1 Sol (max, 55), with Gemini 4 Argon (high) at 50 points; the gap between open-weight models and the leaders widened from 3 points to 12 points, with DeepSeek V4.1 Flash 1017 (max) and GLM-5.3 (max) leading among released weights at 32 points. Per-task costs among the top ten models differ by more than 13 times, with GPT-6.1 Terra (max) lowest at $0.96 and Claude Fable 5.1 (max) highest at $13.04.