Artificial intelligence models and research encompass the foundational algorithms, architectures, and scientific studies that drive system capabilities. Truncated covers new model releases, underlying technical advances, and published research papers.
AI-generated summary Alibaba announces that Qwen4-Max-Preview is now available on Qwen Chat, Agent Mode, and on Alibaba Token Plan. Alibaba states that the model is built for complex coding, office work, front-end design, and professional software tasks.
AI-generated summary Tether AI Research introduces BotPsy as a 0.8B-parameter personal assistant model designed for on-device agentic workflows and function calling. Tether AI Research says BotPsy is built on Qwen3.5-0.8B and achieves tool-use performance for its size while keeping user data private.
AI-generated summary Elon Musk says that Anthropic model Claude Opus 5.5 is a masterpiece and admits his upcoming Grok 4.8 will not match it. Musk stated on X that an expected Grok 5 model should exceed the Anthropic release in two or three months.
AI-generated summary Genspark and Baseten Applied Research announce the post-training of an open-weight model starting from Qwen3.5-397B. The companies state that their new model scores 0.759 on internal evaluations, which outperforms Opus 5.5 and serves as the default production model for Genspark's Deep Research agent.
AI-generated summary Google announces new features and expanded availability for Gemini Intelligence across multiple Android devices and hardware models. Google says the update introduces tools like Rambler and task automation to help users manage everyday tasks.
AI-generated summary NaiveAI has developed an open-weights model called Naive-N0.5-Flash for AI coding tasks, designed to generate a high number of tokens per second. The company states that the model aims to speed up engineering and research workflows to help teams iterate and solve complex problems faster.
AI-generated summary Fish Audio has launched Live 3, which is a duplex voice AI model built for AI agents that operate phone calls. The model can distinguish pauses in phone numbers from the end of a question and has a reaction speed under 200 milliseconds.
AI-generated summary Anthropic has released Fable 5.2 and Mythos 5.2 as new upgrades to its Claude chatbot, alongside safety warnings regarding evaluation challenges and covert capabilities. Anthropic claimed that the new models work 30 per cent faster and more cheaply than their predecessors while handling ambiguous requests better.
AI-generated summary ElevenLabs is launching two new Speech to Text models named Scribe v3 and Scribe v3 Live to improve transcription accuracy and handle messy real-world audio. ElevenLabs says Scribe v3 makes 17 percent fewer diarization errors than Scribe v2 on recordings with three or more speakers.
AI-generated summary The announcement introduces MAI-Transcribe-2 as a transcription model priced at $0.10 per hour of audio. The publisher states that the model is 10x faster than GPT-Transcribe and 5x faster than Gemini 3.5.
AI-generated summary Researchers from Google and the University of Texas at Austin posted a preprint about a model named Muir that reads ten million tokens. The preprint describes how the 1.34-billion-parameter model keeps a working state of 25 megabytes of float32 numbers.
AI-generated summary Unite.AI reports on a research paper proposing a new training framework designed to help Vision-Language Models recognize hidden text in AI-generated optical illusions. The proposed method, named SQUINT, uses knowledge distillation and downscaled image views during training to enable models to read concealed messages more effectively.
AI-generated summary Decision models evaluate input context against predefined possible answers to return a structured decision and probability distribution without generating text. TypeSafe AI introduced Jev and described it as a System One model, while the community refers to these models as decision models.
AI-generated summary Google has added Gemini 4 Carbon Low, Medium, and High variants to the latest Antigravity build according to recent reports. The new variants are not available publicly yet.
AI-generated summary Boost.ai announced the launch of Presis, a self-hosted language model designed to handle classification decisions for AI agent conversations. Boost.ai says internal testing found the model at least 2.5 times faster than the options it replaces.
AI-generated summary Anthropic released Claude Fable 5.2 and Claude Mythos 5.2, models it says show improvements in coding, reasoning, and vision. On CyScenarioBench, Anthropic says Claude Fable 5.2 completed 76.3 percent of challenges compared to 61 percent for OpenAI's GPT-6 Cyber.
AI-generated summary Google is releasing a new Gemini model today according to a report from testingcatalog. The publication states that Gemini 4 Carbon is the most likely candidate alongside a potential Flash model.
AI-generated summary Anthropic tested Claude Fable 5.2 on small coding tasks to observe how prompts about code ownership affect model behavior. Anthropic says the model merged repeated code in about half of the tries when told the code is its own, compared to under a tenth of the tries without that instruction.
AI-generated summary Anthropic has released the Claude Fable 5.2 and Mythos 5.2 models, bringing new artificial intelligence capabilities ahead of an expected market debut later this month. Anthropic says the new models offer faster generation speeds at a lower price, with Fable 5.2 capable of coordinating more than 1,000 agents simultaneously.
AI-generated summary Research concerning how to best prompt or use generative AI and large language models is rapidly aging and becoming obsolete. Users relying on outdated insights for ChatGPT and other popular large language models may find their approaches counterproductive.
The top AI news in your email delivered each morning. Choose categories