This week in AI.

CF

Claude Fable 5.2

Model · Anthropic

5

Anthropic's top-tier Claude model for hard technical work, with cyber and biology safeguards, released November 2026.

Meta's smallest Muse-series language model, positioned as a low-cost option for subtasks and delegation.

I2

Inkling 2 Small

Model · Thinking Machines

2

Open-weights multimodal MoE model (594B total, 17B active) from Thinking Machines, available via Tinker.

L2

LongCat 2.5

Model · Meituan

2

Meituan's open-weights 1.6T-parameter multimodal MoE model (48B active) for agentic coding and visual tasks.

R6

Ripple 65B Preview

Model · Recursive Intelligence

2

Dense 65B open-weight language model from Ricursive Intelligence, trained for agentic coding with up to 1M context.

G4

Gemini 4 Hydrogen

Model · Google

1

Smaller, faster, lower-cost Gemini 4 model from Google DeepMind with a 1M-token context window.

OO

OpenAI o1

Model · OpenAI

1

OpenAI's first reasoning model series, trained with reinforcement learning to think before answering; now deprecated in the API.

Q2

Qwen 2.5 0.5B

Model · Alibaba

1

Small 0.5B-parameter open-weights base language model from Alibaba's Qwen2.5 series, Apache 2.0 licensed.

Ant Group's multimodal mixture-of-experts model (560B total, 25B active) in the Ling family, served via API.

E5

ERNIE 5.5 Flash

Model · Baidu

1

Baidu's open-weights 280B-parameter sparse MoE base model (14B active) with 1M-token context, released pre-post-training.

MH

MiniMax H3 Image

Model · MiniMax

3

Open-weights MiniMax image generation and editing model producing up to ~16 MP images in 6 sampling steps.

S5

Seedream 5.5

Model · ByteDance

1

ByteDance Seed's text-to-image and image-editing model, part of the Seedream family.

PJ

PJiT

Model · Linum AI

1

2B-parameter pixel-space text-to-image diffusion model from Linum AI, trained from scratch for about $21K; Apache 2.0.

N2

NitroFusion 2

Model

1

One- and two-step image generation model distilled from Z-Image-Turbo using text-stream adversarial distillation.

Custom-built visual interface for leaving timestamped video feedback that an AI agent then applies, Frame.io-style.

HI

Higgsfield

App · Higgsfield

1

AI creative suite for generating images, video and voice, with an MCP connector for use inside Claude.

DR

DreamX

Model

1

Research model that turns a pretrained video diffusion model into a latent action learner for human-to-robot motion transfer.

RE

RelayVSR

Model

1

Streaming real-world video super-resolution using sparse large-model keyframes and a lightweight propagation network.

Vocal synthesis tool reportedly used by an AI model to generate lead vocals, harmonies and chops for a song.

MO

MOSS-Live

Model

1

Full-duplex real-time voice, video and screen assistant from MOSS (模思智能) with background task execution.

VO

VoxLution

Model

1

Research model generating 3D indoor scenes as sparse voxels, refined from 25 cm to 1.5625 cm by an autoregressive cascade.

3D reconstruction model that recovers geometry and camera poses from photos or video using a loop transformer.

MO

Mobius

Model

1

Unified motion-language model for text-to-motion, motion captioning, completion and editing, from HKUST and Tencent Games.

HA

HairDeck

Model

1

Research model that generates animation-ready stylized hair cards and spring-bone chains from a single character image.

FL

FlowHMR

Model

1

Monocular video motion capture that combines flow matching with physics-based RL for plausible 3D human motion.

RE

ReLift

Model · Nvidia

1

Feed-forward model that repairs lifted 2D tracks and depth into accurate dense 3D trajectories from monocular video.

LL

Lucid Lyra

Model · Nvidia

1

NVIDIA research model that turns a single image into a queryable 3D Gaussian scene with semantics and object instances.

CO

CoLa3D

Model

1

Research model that decomposes a single image or scene mesh into object-level 3D meshes with consistent spatial layout.

AC

Adobe Creative Cloud AI Agents

Agent · Adobe

1

Adobe's AI agents and assistants across Creative Cloud apps, announced at Adobe MAX 2026, plus Firefly design layouts.

Personal scheduled AI bots that email recurring reports, shown in Matt Wolfe's "AI operating system" and Bass Coach demo.

User-built agent skill that checks decisions against a personal one-page doc of goals, energy drains and priorities.

A scheduled AI workflow that reads email, calendar and Slack and returns a numbered list of suggested actions.

An agent skill that updates other skills by reviewing many recent conversations for recurring feedback.

CH

ChatGPT

App · OpenAI

1

OpenAI's chat assistant for answering questions, writing, image generation, coding and agent-style tasks, on web and apps.

SP

Spacefast

App · Automattic

1

Automattic's hosting platform for deploying vibe-coded apps, with built-in database, auth, email and analytics.

SF

SWE-2 Flash

Model · Cognition

1

Cognition's fast, efficient coding model, built for sub-agents handling delegated coding tasks.

BI

Businesses in ChatGPT

App · OpenAI

1

OpenAI's beta Business Hub for claiming business profiles in ChatGPT, viewing engagement insights, and launching Business Agents.

CM

Claude Meetings

App · Anthropic

1

Anthropic's Claude feature for joining meetings, taking notes and working on a shared whiteboard.

CW

Claude Whiteboards

App · Anthropic

1

Collaborative digital whiteboard artifact in Claude for laying out ideas in meetings and having Claude summarize them.

CI

Claude Interactive Connectors

App · Anthropic

1

Anthropic's MCP Apps-based feature that opens connected tools beside or full-screen in Claude chat, where Claude can edit them.

CD

Claude Data

App · Anthropic

1

Claude feature that connects Claude to your data so you can ask questions and analyze it, available on paid plans.

CS

Claude Sheets

App · Anthropic

1

Spreadsheet feature inside Claude that builds and edits sheets with tabs, formulas and drop-downs from a prompt (beta).

Agent skill that researches a podcast guest and topic, suggests title and thumbnail angles, and drafts an interview guide.

ZA

ZAP

Model

1

27B vision-language robot policy trained only in simulation, run zero-shot on a real bimanual robot.

Open-source robotics toolkit of four libraries: motion planning, motion retargeting, grasping and agent skills.

AM

Adobe MAX Sneaks

Research · Adobe

1

Adobe research preview that generates and syncs sound effects for video clips, with a 2D canvas for blending sounds.

Tiny ~51M-parameter vision model from Cactus Compute for on-device object detection, text reading and spatial understanding.

ST

STRATA

Model · Nvidia

1

Nvidia's 332M-parameter AI model emulating the SCREAM global storm-resolving atmosphere simulation at ~4.9 km, 10-minute steps.

Google's API for AI-driven decision-making, reported by creators as a counterpart to similar OpenAI and Microsoft offerings.