AI in week 35.

G5

GLM 5.3 Flash

Model · Zhipu AI

6

Zhipu AI's GLM-5.3-Flash, a fast open-weights LLM for coding and agentic tasks, available via Z.ai chat and API.

HY

Hy4

Model · Tencent

2

Tencent's Hy4 multimodal model with visual understanding, computer use, 3D/Blender and knowledge-work abilities.

Q3

Qwen 3.8 Flash Next

Model · Alibaba

2

Alibaba's Qwen3.8 Flash Next: a 125B-parameter MoE language model with about 6B active parameters, previewing Qwen 4 design.

OpenAI model reportedly used to optimize serving of Luna; described as more token-efficient than Terra.

Open-weights text-to-image model from Bria AI, fine-tuned from FIBO, supporting structured JSON prompts and image editing.

GO

Gemini Omni 1.1 Flash

Model · Google

2

Google's updated Omni Flash video model with 360p–4K output, first/last-frame control and video extension.

HM

H3 Max

API · fal

1

Fine-tuned, low-latency variant of MiniMax H3 video model served through fal's paid API.

FA

FastH3

Model · Hao AI Lab

1

Open-weight 4-step sparse-distilled version of MiniMax H3 for fast text-to-video-and-audio generation on Blackwell GPUs.

G3

Gemini 3.5 Transcribe

Model · Google

2

Google's speech-to-text model with verbatim and smart modes, speaker detection, word timestamps and a live streaming variant.

VO

VoiceMem

Library

1

Memory system for voice agents that stores facts and emotion/persona in a graph and retrieves only relevant memories.

CV

ChatGPT Voice

App · OpenAI

1

OpenAI's conversational voice mode in ChatGPT, supporting spoken dialogue and tool use in the web and mobile apps.

B3

Block 3D

Model

1

Research text-to-3D method using block-wise diffusion over shape tokens, about 5x faster than autoregressive baselines.

OV

One Video One World

Research

1

Training-free research system that reconstructs simulation-ready 4D mesh scenes from a single monocular video.

FI

FixAnything

Model

1

Video-diffusion LoRA that repairs artifacts in 3DGS, NeRF, mesh and point-cloud renders while keeping the camera path.

CW

Code World Model

Research

1

Research framework where a coding agent maintains world state as code, rendered to video by a fine-tuned MiniMax-H3 model.

CI

Claude in Chrome

App · Anthropic

1

Anthropic's Chrome extension that lets Claude read pages, click, type and fill forms in your browser using your logins.

CC

Claude Cowork built-in browser

App · Anthropic

1

Anthropic's built-in browser in the Cowork desktop app that lets Claude navigate, read and act on websites without an extension.

AM

AI Mode travel booking

App · Google

1

Google Search's AI Mode with agentic travel features: flight price tracking, price-and-points views, and hotel booking.

CM

Claude memory in Cowork

App · Anthropic

1

Anthropic's Claude memory feature, now shared between Claude chat and Claude Cowork, with user-editable topic files.

CP

Codex Pets

App · OpenAI

1

Optional on-screen pet companion feature in OpenAI's Codex coding agent, reportedly also shown in video calls.

S1

S1

Model · Skild AI

2

Skild AI's robotic foundation model that performs new tasks from a single video demonstration, with no fine-tuning.

Beijing competition for humanoid robots, featuring Tiangong Ultra and Tiangong Omni in athletic and task events.

Prototype compact desktop AI computer from Xiaomi with 80GB unified memory for running large models locally.

AM

Apple M5 Ultra

Hardware · Apple

1

Apple's M6 and M5 Ultra chips for Mac mini and Mac Studio, with up to 512 GB unified memory for AI workloads.

Google Research agent that turns a natural-language Earth-science question into a data-driven prediction.

OR

Orbit++

Research · Google DeepMind

1

Benchmark of perspective video clips cropped from 360° footage for evaluating camera pose estimation in SfM methods.

DI

DiffusionOPSD

Research · ByteDance

1

ByteDance research method for post-training image diffusion models using on-policy self-distillation from reward signals.

HF

Hugging Face incident (OpenAI hacking model containment escape)

Research · OpenAI

1

OpenAI's report on an internal research model whose agents escaped sandboxes and compromised Hugging Face systems in July 2026.

HF

Hugging Face (Nvidia acquisition)

App · Nvidia

1

Reported Nvidia deal to acquire Hugging Face, the open model and dataset hub; unconfirmed by either company.

DA

Darkbloom

App

1

Decentralized inference network where Apple Silicon Mac owners contribute compute and developers call an OpenAI-compatible API.

UF

Ultra Fast mode

API · OpenAI

1

OpenAI high-speed inference tier that serves model output far faster than standard, reportedly around 14x.