ai-release

Canva 推出了其首个视觉语言模型 Brush，能够理解设计的布局、图层与风格，并协助用户进行文本修改、整套演示文稿重设计以及从空白页面生成设计。

Elon Musk quote-tweets Sundar Pichai's announcement that Gemini 4 Argon is available globally for developers and enterprises via the Gemini API, as well as for subscribers in Google Antigravity, the Gemini app, and Search, praising direct CEO communication on X.

A major frontier AI lab has unveiled a new flagship model claimed to be the most powerful in the world, representing a significant step forward in AI capability.

Google's Sundar Pichai announced an early look at Gemini 4 Argon, claiming frontier performance across complex workflows, cyber defense, and software engineering, with heavy internal use at Google; the post is shared as a humorous 'don't be benchmaxxed' reaction.

Liquid AI announced d1, their first decision model, which outperforms Jev on Hugging Face's Decision Index with advantages in multilingual evaluations, robustness, and handling longer inputs.

Opus 5.5 is released, with three.js highlighted as the biggest winner from the update.

Alex Finn shares his excitement about early access to Grok Bot for Tesla, demonstrating its capabilities in a video ride in his Cybertruck.

Anthropic is set to release Sonnet 5.5 and Haiku 5.5, which are upcoming updates to their AI language models.

Anthropic has released Claude Opus 5.5, featuring price drops, increased five-hour usage limits for Pro, Max, and Team plans, and a new rate limit reset option for subscribers.

The post suggests that GPT-6 is already outdated with the release of SHIP 6.1, implying a new AI model or software update.

Introduction of Jesse v1, a new AI model architecture built from scratch using public datasets, consuming zero tokens and offering fast reasoning capabilities.

Figure releases Helix 2.5, an AI model that enables robots to generalize household tasks zero-shot across 30 homes without additional training.

Apple has released their Apple Foundation Models (AFM) for local use on MacOS 27, accessible via a terminal command, prompting discussions on the impact of local AI and open models.

DeepSeek has released V4.1 Flash, a 552B MoE model with efficient active parameters, achieving performance close to GPT-5.6 Sol at a much lower cost.

This article likely covers the release of LLM version 0.35, which is an update to a large language model.

Abhinav Anand, a 20-year-old from Bihar, has released Arcle V1, an open-weights multimodal AI model with 5.84 billion parameters, trained independently without a team or investors.

Nativ v0.3.7 is released, introducing features like projects for grouping chats, import/export functionality, new speech models, and enhanced local AI capabilities for Apple Silicon Macs.

Sam Altman reassures users that OpenAI is working to make Astra available to everyone quickly, acknowledging frustration and asking for patience.

Release of a speculative decoding implementation in Uzu, initially supporting Qwen3.6 27B with upcoming support for Qwen3.8 27B and Muse Glimmer.

IFM AI introduces K2 Horizon, a fully open-source fleet of six foundation models ranging from 0.9B to 375B parameters, claiming frontier performance in coding and agentic tasks. The release is touted as the largest fully open-source model launch in AI history, with complete transparency and open access to training data and code.