Models & Research · Fri 6 Nov

AIAI-generated summary MiniMax announced that the M3.1 Flash model is available on FireworksAI. MiniMax says representatives from MiniMax and FireworksAI will discuss building the model for long-running agents at an upcoming event in San Francisco.

AIAI-generated summary Apple released the watchOS 27.3 beta this week alongside iOS 27.3, introducing features such as Siri Recap and Live Rewind for recent Apple Watch models. Apple says Siri Recap ambiently takes notes on daily conversations to generate summaries, while including privacy protections that prevent audio storage and sensitive data logging.

AIAI-generated summary MIT researchers and BASF scientists developed CheMeleon, an open-source molecular foundation model designed to predict chemical properties when task-specific experimental data are scarce. William Green says the model reduces the number of molecules that must be experimentally tested to accurately predict new compounds properties.

AIAI-generated summary Physical Superintelligence announced on November 6, 2026 that an internal research system running instances of a non-public Anthropic model found a proof of the Aaronson-Ambainis conjecture. Physical Superintelligence says the system produced a complete proof during a run on its infrastructure, and the Lean proof assistant accepted it.

AIAI-generated summary Cerebras noted that Moonshot released the Kimi K3.1 Pro model. Cerebras stated that it used the model to complete an online shopping task in 11.5 seconds.

AIAI-generated summary Nace AI has released a report detailing the training process for Drex, which is a family of open-weight AI models designed to select the most likely answer from a fixed list. The model called Drex v1.5 was trained using verified answers for tasks including contract reading and phishing detection, and it can be tested within Nace Console.

AIAI-generated summary Meta says everyone using its Muse agent and Meta AI now gets new models trained using its frontier model Muse Wonder. The new models include two features named Active Memory and Active Focus that were introduced with Wonder.

AIAI-generated summary Meta AI says specialized models trained using Muse Wonder are coming to @Muse and Meta AI for complex tasks. Meta AI says the updates include Active Memory and Active Focus to help users stay aligned with goals.

AIAI-generated summary Thinking Machines Lab released Inkling 2 Small, a model with 594 billion parameters that will have its weights released next week. Artificial Analysis stated that the model scored a 31 on the Artificial Analysis Intelligence Index and noted its pricing is somewhat expensive.

AIAI-generated summary Salesforce introduced Koa Decisions on November 6, 2026, as a decision model built from Google’s open-weight Gemma that accepts text, images, and audio. Salesforce plans to expand the technology beyond Agentforce request routing into Salesforce workflows and developer tools.

AIAI-generated summary A new paper published on the arXiv preprint server argues that artificial intelligence models are trained on the traces of thought rather than actual thinking processes. The authors state that this focus on final traces hobbles current artificial intelligence systems and can create similar problems for human users relying on them.

AIAI-generated summary Five scientific experts discuss whether artificial intelligence is or might one day become truly intelligent. The experts from various fields offer different answers regarding the vagueness of the term intelligence.

AIAI-generated summary Google is developing an offline mode for Gemini Live that requires an active data connection for the initial setup process. Code discovered in the latest Google app indicates that the offline experience will feature a reduced set of capabilities and limited features.

AIAI-generated summary Thinking Machines has unveiled Inkling 2 Small as its second model generation, featuring 17 billion active parameters out of a total 594 billion. Thinking Machines says the model is strongest in text recognition, infographic interpretation, and visual reasoning.

AIAI-generated summary A teaser titled The becoming begins has been posted on the Qwen X account. The source expects new models including Qwen4-Max, Qwen4-Flash, Qwen4-Plus, and Qwen4-27B to arrive soon.

AIAI-generated summary Thinking Machines released Inkling 2 Small, a 594B-total and 17B-active multimodal MoE trained in 68 hours on 12T tokens. The model features a 1M context window and is available on Tinker with Apache 2.0 weights promised next week.