China AI

AIBAAI has released AREX-Base, a 122B-total, 10B-activated Mixture-of-Experts deep research agent built on Qwen3.5-122B-A10B with a 262,144-token context.

Why it matters: The release pairs a 122B-parameter deep research agent with benchmark tables against frontier and open models, letting readers compare its search-agent results directly.

AIBAAI's AREX-Turbo is a dense 4B deep research agent built on Qwen3.5-4B with a 262,144-token context length.

AIAwni Hannun argues that Apple pays Google about a billion dollars per year for Gemini, even though open models such as GLM 5.2 and the upcoming Kimi K3 could be used for free.

AIIf open labs keep finding 2.5x a year, compute advantages depreciate fast. The frontier isn't who has the most FLOPs. It's who converts them best. 👇

AIA Chinese company has revealed an autonomous robot toilet that drives itself to users when summoned by voice command or remote control. The post provides no further technical details, pricing, or availability.

AIKimi has temporarily paused new subscriptions after Kimi K3 demand pushed its GPU capacity close to limits over the past 48 hours.

AIIntern Large Models releases Intern-S2-Preview-397B, a 397B-parameter open multimodal foundation model for scientific intelligence and long-horizon agents.

AIThe Chinese Academy of Sciences released ScienceOne Omni 2.0 (Pan Shi 2.0) at the World Artificial Intelligence Conference on July 17.

AIMoonshot AI announced Kimi K3, a native multimodal model with 2.8 trillion parameters and a 1 million token context window.

AIByteDance's UniVR-34B-Planning, built on Emu3.5 at 34B parameters, learns visual reasoning, physical dynamics, and long-term planning from visual demonstrations using a next-token objective and two-stage training on the VR-X dataset with VR-GRPO reinforcement learning.

AIIn a long interview, Peking University professor Dong Bin describes his team's AI framework autonomously solving the Anderson conjecture, reportedly the first such domestic result with large-scale formal verification.

AIUBTECH and Hitachi (China) signed a strategic cooperation agreement to jointly develop humanoid robot solutions for smart manufacturing across fields including elevators, building systems, healthcare, and semiconductor equipment.

AIUBTECH's exclusive distributor for Germany, Switzerland and Austria, Terra Robotics, has bought several dozen full-size humanoid robots that are being deployed at logistics centers of ROSSMANN, a European health and beauty retailer with annual turnover above €15 billion.

AIThe Hong Kong Stock Exchange named UBTECH Robotics (09880.HK) a constituent of its inaugural HKEX Tech 100 Index on December 9.

AIBAAI introduced RoboBrain Orca, which learns world latent representations from multimodal data and models world-state transitions through Next-State-Prediction instead of only next-token, next-frame, or next-action prediction.

AICognition says its SWE-1.7 model, developed from the open-source Kimi K2.7 Code base, performs as well as or better than leading U.S. frontier models on its new trustworthiness evaluation suite.

AISkywork users can now set custom slide sizes instead of being limited to 16:9. The update supports portrait decks, mobile-first presentations, social posts, and reports in the formats users need.

AIPromptArmor reports that Skills in Microsoft Copilot Cowork can call DeepSeek even when an organization has not opted into the DeepSeek Preview.

AIZ.ai has launched ZCode, its official development environment for GLM-5.2, available on macOS, Windows, and Linux.

AIXiaomi MiMo's account celebrated growing developer adoption of its open-weights models, crediting Cline for building on MiMo.

AIPaddleFormers 1.2 is released with support for training DeepSeek-V4 and 128K+ long-context training.

AISebastian Raschka's tutorial shows how to build a fully local coding agent by pairing an open-weight LLM served through an inference runtime with an open-source harness that can read files, edit code, and run commands.

AIQwen released Qwen3-ForcedAligner-0.6B-hf, a Transformers-format forced aligner that predicts timestamps for arbitrary units within up to 5 minutes of speech in 11 languages.

AIThe AI Futures Project forecasts that China will reach commercial-scale 7 nm-capable immersion DUV around the mid-2030s and commercial-scale 5 nm-capable EUV around 2040.

AIXiaomi has launched MiMo Claw, an agent built on its flagship MiMo model and integrated with Kingsoft Office for Word, Excel, PowerPoint, and PDF workflows.

AIBAAI President Wang Zhongyuan announced a shift in AI from token prediction to physical state prediction in the institute's 2026 annual research report.

AIZ.ai has released GLM-5.2, its flagship model for long-horizon tasks, which it says substantially improves on GLM-5.1 and supports a 1M-token context.

Why it matters: The source gives benchmark tables against named rival models and deployment settings, useful for judging where GLM-5.2 sits among current flagship models.

AIZ.ai's release notes list GLM-5.2 as supporting 1M lossless context, with improved long-horizon task performance and reduced context drift and goal forgetting.

Why it matters: The page lists a dated series of Z.ai model releases, showing how the coding and long-horizon agent line has evolved from GLM-4.5 through GLM-5.2.

AIByteDance has released Sa2VA-LLaVA-1.5-7B on Hugging Face, a model built on LLaVA-1.5-7B with a SAM2 grounding encoder that performs dense image and video referring segmentation alongside open-ended chat.

AITuring Award winners Whitfield Diffie and Andrew Barto delivered keynotes at the BAAI Conference on AI security and reinforcement learning.

AITri Dao's post proposes loading SSM states, computing with them, and not storing them to make hybrid model decoding about 2x faster.

AIMoonshot AI released Kimi K3 on Hugging Face as an open-weight, native multimodal agentic model with 2.8T total parameters and 104B activated parameters.

Why it matters: The release pairs open weights with a 2.8T-parameter MoE architecture and benchmark tables against several named closed models, useful for comparing frontier capability claims.

AIOpenRouter introduced Fusion, a tool that sends a prompt to a panel of models and has a judge model fuse their results into one answer.

Why it matters: The source gives benchmark scores, panel compositions, and contamination controls, letting readers judge how much of the gain comes from model diversity versus self-synthesis.

AIMoonshot AI published Kimi-K2.7-Code, a coding-focused agentic model built on Kimi K2.6, with a 1T-parameter MoE architecture and 32B activated parameters.

Why it matters: The model card gives benchmark comparisons against GPT-5.5 and Claude Opus 4.8 on coding and agentic tasks, useful for judging its position among current coding models.

AIXiaomi MiMo has released MiMo Code V0.1, an open-source AI coding assistant for the terminal under the MIT license.

Why it matters: The post specifies MiMo Code's Claude Code compatibility and MIT license, which bear directly on whether existing coding-agent setups can migrate without rework.

AIByteDance open-sourced inference code and model weights for Bernini-Diffusers, a full video generation and editing pipeline with an MLLM-based semantic planner and a DiT-based renderer.

AIEvoQuality is a ByteDance vision-language model for no-reference image quality assessment that generates pseudo-ranking labels through pairwise majority voting and refines them with GRPO, requiring no human-annotated quality scores.

AIByteDance's Sa2VA-Qwen3-VL-4B-SAM3 is built on Qwen3-VL-4B-Instruct with a SAM3 grounding encoder and produces dense image and video referring segmentation alongside chat.

AIZ.ai released SCAIL-2, an open-source model that animates a reference character from a driving video without skeleton maps or inpainting masks.

AIMiniMax describes MaxProof, an evolutionary-search framework that lets its M3 model refine candidate proofs over multiple rounds.