AI in week 43.

KK

Kimi K3.1

Model · Moonshot AI

7

Moonshot AI's open-weight 2.8T-parameter multimodal MoE agentic model with 1M-token context.

SA

Sage A1.5

Model · Alpha Innovator

1

35B mixture-of-experts open-weights agentic model for engineering, long-horizon search and scientific research.

CA

Carbon-X

Model · Hugging Face

1

Open 14.8B language model from Hugging Face that reads raw DNA in chat and answers in English, built on Qwen3-14B.

L3

Ling 3.0 Flash Santé

Model · inclusionAI

1

Open-weights 124B MoE healthcare model (5.1B active) from inclusionAI, specialized from Ling-3.0-flash.

M1

MemT 1.0

Model · MemTensor

1

Open-weights memory foundation model from MemTensor with a fixed-size persistent memory state, built on Qwen3.6-35B-A3B.

MM

MiniMax M3.1 Flash

Model · MiniMax

1

MiniMax's fast, low-cost language model aimed at agentic coding and automation, available via API.

PO

Perplexity open-source decision models

Model · Perplexity

1

Two open-sourced Perplexity decision models, at 4B and 27B parameters, announced by the company.

GP

GPT-OSS

Model · OpenAI

1

OpenAI's open-weight reasoning models, gpt-oss-120b and gpt-oss-20b, released under Apache 2.0.

IV

Ideogram Vector

Model · Ideogram

2

Ideogram's model for generating editable SVG vector graphics from text prompts or by converting existing images.

WE

WeVisOne-VL

Model · Tencent

1

Tencent's Qwen3-VL-based model unifying detection, pointing, OCR, segmentation and alpha matting in one VLM.

Bria's open-weights image generation and natural-language editing model with multi-reference support, non-commercial license.

GI

GPT Image 3 (Aurora)

Model · OpenAI

1

Unconfirmed OpenAI image-generation models, codenamed Aurora, spotted in API references and arena tests in October 2026.

CC

Canva Craft

Model · Canva

1

Canva's first in-house image generation model, a two-stage planner-generator trained on designs, output up to 2K.

Community LoRA for the MOSS Image model that applies an ink-splatter, sumi-ink art style to generated images.

8-step distillation LoRA for MOSS-Image-SFT that runs without CFG, for ComfyUI.

Community full fine-tune of MOSS-Image-Base at 1024 px, with optional 8-step turbo LoRAs; open weights under Apache-2.0.

FO

FastH3 Omni

Model · Hao AI Lab

1

Distilled MiniMax H3 reference-to-video-and-audio model from Hao AI Lab, generating in 8 steps instead of 50.

EC

ElevenLabs Chimera

App · ElevenLabs

1

ElevenLabs tool that lip-syncs a still image to an audio file, using Seedance 2.5 and Eleven v4.

O2

OmniHuman 2.0

Model · ByteDance

1

ByteDance image-and-audio-driven digital human video model, offered through the BytePlus API.

HK

Higgsfield Katana

App · Higgsfield

1

Higgsfield tool that recreates viral videos shot by shot with your own characters, products or footage, used inside Claude.

M3

Ming 3.0 TTS

Model · Ant Group

1

Open-weights text-to-speech models from Ant Group's Inclusion AI with voice design, cloning and tag-based control.

DE

Dessa

Model · Block

1

Block research preview of a full-duplex speech-to-speech model trained only on human conversation audio, without text.

BA

Bark3

Model

1

Compact English speech-to-text model family from Underdog that runs on-device on Mac, starting at 163 MB.

V3

Voxtral 3

Model · Mistral

1

Mistral's voice model for speech agents, with real-time diarization, interruption handling and tool calls.

H1

Hi3D 1.0

Model · Hi3D

1

Open-source image-to-3D and text-to-3D generation model from Hi3D, formerly Sparc3D, trained on ~2M curated 3D assets.

HY

Hy-Nomas

Model · Tencent

1

Tencent Hunyuan interactive world model generating real-time, controllable 720p video with memory for consistency.

PU

PURE

Research

1

Post-training method that refines feed-forward 3D reconstruction models (Pi3, DA3, VGGT) on unlabeled video.

SK

Sky-001

Model · Skyfall

1

Engineering world model from Skyfall that decomposes mechanical objects into CAD parts and generates floor plans.

JS

JST-2

Model

1

Viggle's physics-native world foundation model that generates 3D scenes, rig-free characters and motion from text or images.

T3

T3lescope

Research · Preferred Networks

1

Generative method from Preferred Networks that reconstructs high-resolution 3D scene meshes from posed multi-view images.

Unreleased, unnamed ByteDance world model demoed in Beijing for novel camera paths and text-to-world generation.

WO

WorldGacha

App

1

Free browser tool that procedurally generates fictional regions and towns as manga-style background line art.

CD

Claude Duet

Agent · Anthropic

2

Anthropic's proactive agent app that monitors connected apps, sends briefs and completes tasks on its own virtual machine.

OR

OpenAI research agents

Research · OpenAI

1

OpenAI, PerturbAI and Scripps Research run thousands of GPT-Rosalind agents in parallel to analyze brain-cell gene data.

YA

Yelp AI restaurant calls

Agent · Yelp

1

Yelp app feature where an AI voice agent phones a restaurant to check wait times and texts the result to the user.

QC

Qwen Coder Next

Model · Alibaba

1

Alibaba's Qwen coding-specialist language model, available as open weights for local and hosted use.

CL

ChatGPT Locked Chats

App · OpenAI

1

ChatGPT feature that moves selected chats into a PIN- or biometric-protected section with its own separate memory.

UH

Unitree H2X

Hardware · Unitree

1

Unitree H2X is a full-size humanoid robot from Unitree, shown in service and guard-robot demos; specs are largely undisclosed.

UH

Unitree H2 Kung Fu Legend Guardian

Hardware · Unitree

1

Full-size humanoid robot from Unitree Robotics, shown in demo videos performing flips, kicks and martial-arts moves.

MS

Mac Studio

Hardware · Apple

1

Apple's compact desktop workstation with M5 Max or M5 Ultra chips and up to 512GB unified memory for local AI and pro work.

Unreleased OpenAI consumer AI device; no specifications, price or launch date have been published.

OC

OpenAI custom AI chip

Hardware · OpenAI

1

OpenAI's in-house AI accelerator chip, developed to lower the cost of training and running its models; not publicly available.

PV

PHP vulnerability found by GPT-5.6

Research · OpenAI

1

AI-assisted discovery of two critical, roughly 20-year-old PHP remote code execution flaws, dubbed PHP2Shell.

MM

Microsoft MAI-DxO

Model · Microsoft

1

Microsoft AI's multi-agent diagnostic orchestrator for complex medical cases, announced as a step toward medical superintelligence.

OM

OpenAI math proofs / AI scientific discovery results

Research · OpenAI

1

OpenAI's reported math proofs and scientific discovery results produced by internal, unreleased models.

DE

DeepInfer

Library · DeepSeek

1

Open-source inference framework from DeepSeek for serving DeepSeek models on NVIDIA SM100 GPUs and Huawei Ascend 950 NPUs.

ZD

Zoom deepfake risk detection

App · Zoom

1

Zoom feature that analyzes video and audio in real time to flag possible deepfake participants, with a confidence indicator.