AI takes shape, intelligence reborn

OpenAI's sixth-generation flagship model released 2026/9/3 (PT): billed as the world's most intelligent and aligned model, built for long-horizon agentic work and computer use; first model to reach the Critical cybersecurity threshold; 99.9% ARC-AGI-3, 97.6% FrontierMath Tier 4, 100% ExploitBench; API $10/$50

ByteDance flagship released 2026/6/23, rivaling GPT-5.5/Opus 4.7; 10+ hour autonomous engineering, 2-hour video understanding, 80% lower cost than Opus

AI assistant by xAI, real-time web, integrated with X/Twitter, witty style

Microsoft AI assistant with Bing search and M365 ecosystem

AI role-playing chat platform with various virtual characters

All-in-one AI chat by Quora, aggregating GPT/Claude/Gemini and more

AI assistant by Kunlun Tech, DeepResearch + search, popular Chinese AI

AI search and learning tool by Alibaba, top Chinese AI search traffic

Multi-agent AI search by 360, DeepSeek integrated, generates videos/reports/PPTs

AI assistant by Meta, Llama-based, across WhatsApp/Instagram/Facebook

AI-native browser by Perplexity with built-in AI assistant for emails, shopping, and scheduling

Xiaomi ecosystem AI assistant, 2024/10 launch with 2026/4 V7.12 adding memory/schedule/translation/research/discovery; covers 600M devices

Google $99 smart speaker released 2026/6/18, first hardware built for Gemini; natural multi-step commands and continuous dialogue

Meta x EssilorLuxottica AI smart glasses from $299, with native Muse Spark multimodal model; 26 styles, targeting 10M units in 2026 H2

Tencent WeChat AI assistant launched 2026/6/20 (v8.0.75), first-level entry at top-left; voice commands for calls, mini-programs, playlists; 1.4B user Agent entry

A near-real-time voice conversation model released by Google DeepMind on 2026/9/15, with two variants: the standard version and Live Extended Thinking. The latter reasons while it speaks and targets complex multi-step tasks, ranking first on the Speech to Speech Quality Index at 82.6, scoring 68.6% on τ-Voice and 97.7% on Big Bench Audio. It supports mid-conversation switching across 97 languages, user interruptions, background tool calls and asynchronous long-running tasks, and embeds SynthID watermarks in its audio.

xAI's frontier model for coding, agentic tasks and knowledge work, 500K-token context, $2 input and $6 output per million tokens, AA-Briefcase score of 1657, with a fast variant limited to Cursor and Grok Build

Anthropic's first Claude 5.5 model, released 2026/9/22: matches Claude Fable 5.1 on most work; $4 per million input tokens and $20 per million output, cache reads down 60%, total operating cost about 40% below Opus 5 and output more than 30% faster

OpenAI light-tier GPT-6 model released 2026/9/22 (PT) for high-volume jobs with a clear goal such as summarization, extraction and short answers; API priced at $0.10 per million input tokens and $0.50 per million output, half of GPT-5.6 Luna, available to Free and Go users in the desktop app

OpenAI's mid-tier model launched 2026/10/27 at $1 per million input tokens and $5 output with $0.05 cached input, half the rates of GPT-6.1 Sol, built for computer use and reported to beat GPT-6 Astra on vision at about 47 percent the latency and 92 percent less cost

Moonshot K3 post-train: 2.8T params, AA index 55, 2 dollars per M input tokens

DeepSeek new-generation flagship model with improved reasoning, coding, and multilingual capabilities

Meituan's new large model preview launched on 2026/9/25, continuing the MoE architecture with native image understanding and deep adaptation to Claude Code

NVIDIA multimodal AI agent model, efficient inference for edge and robotics

FaceBrain open-source multimodal model, vision/audio/text, efficient on-device deployment

DeepSeek's smallest new-architecture model, launched Sep 10, 2026 and refreshed to the 1017 checkpoint on Oct 17: 552B backbone plus 196B Engram (about 763B total weights), 8B/16B active params, native vision, 1M context, FP4 KV cache at ~890 bytes/token; the 1017 build scores 77.0 on DeepSWE v1.1, MIT-licensed with API input from 1 yuan per million tokens

Chat assistant by Mistral AI, Europe's top open-source, strong multilingual

AI assistant by Inflection AI, emotional companionship and daily chat

AI search engine, multi-model, privacy-first smart search

AI assistant with free unlimited DeepSeek R1 and web search

Multi-functional AI: chat, image gen, writing, 13M monthly visits

Multi-agent AI search workspace, all-in-one AI assistant

AI chat by Baichuan Intelligence, strong Chinese comprehension

Multimodal AI by StepStar, first-tier Chinese model for chat and creation

AI Q&A by MiniMax (Hailuo AI parent), leading product and traffic

Large model chat platform by Meituan, outstanding Chinese dialogue

AI assistant by Huawei, DeepSeek integrated, HarmonyOS ecosystem

Flagship code and Agent model launched by Zhipu, with significantly enhanced post-training capabilities; weights are temporarily not open-sourced due to security hardening strategies

Google's new-generation model in the Flash series released on 2026/9/2 (the third Flash version within six weeks), focusing on long-horizon coding agents and professional workflows: DeepSWE v1.1 73.7%, Terminal-Bench 2.1 89.4%, Vals Finance Agent v2 61.4%; 1M context/64K output, intro price $0.75/$3.75 per million tokens; also launched Flash Cyber for defensive side at the same event

Tencent Hunyuan, a new generation flagship large model released and open-sourced on August 28, 2026 (Apache 2.0), with 770B total parameters, 49B activated, and 1M context, positioned as 'built for productivity'; internally, 163 expert engineering blind tests averaged 2.99/4, covering software engineering, office analysis, game development, and scientific research scenarios, with WorkBuddy, Yuanbao, and ima launched simultaneously

Meta's personal AI agent. On 2026/9/17 it released Muse for Mac, which executes tasks directly on the local computer once the user grants explicit authorization, including tidying the Downloads folder, finding lost files and summarising messages and notes; launched in the US on 2026/9/8 as a chat interface backed by a private cloud VM whose visible browser shops, books tickets, sends emails and fills forms for you, with always-on long-running tasks and cross-app memory protected by the MuseSecureVM and Sentinel architecture, free tier plus $20 and $100 monthly plans

OpenAI's full-duplex speech model on the API (Sep 10, 2026): a single model that listens and speaks simultaneously, responds to interruptions in real time, and delegates deep reasoning and tool calls to backends like GPT-6 Astra, at $0.05/min for the voice layer

Moonshot AI renamed Kimi WebBridge to the Kimi browser extension on September 22, 2026, adding a sidebar to navigate, fill forms and complete multi-step web tasks via chat, with repeated actions recordable as replayable skills; it drives the browser locally over the Chrome DevTools Protocol so cookies and login sessions never leave the machine, and works alongside Claude Code, Cursor, Codex and other agents

A mid-tier model released by Anthropic on September 28, 2026, with output speed more than 30% faster than Sonnet 5 and per-task cost reduced by up to 30%, while pricing remains unchanged.

A model released by OpenAI at DevDay on September 29, 2026, succeeding GPT-6 Sol, which had been online for only 7 days. Its capabilities are close to GPT-6 Astra, while its price is about one-fifth of it.

A persistent agent released by OpenAI on September 29, 2026, powered by GPT-6 Astra. Each dot has its own cloud computer and can work continuously 24/7 across more than 4,000 apps.

A frontier model released by Google DeepMind on September 30, 2026, the first in the Gemini 4 series; the output token limit has been increased from 64K to 1 million, with a focus on long-horizon software engineering, enterprise knowledge work such as law and finance, cybersecurity defense, and creative writing; it is first made available to trusted cyber defenders through the Fairwind program, and subsequently to paid API customers and Google AI Ultra subscribers

Anthropic released this small model on October 7, 2026, calling it the cheapest and fastest model it has shipped. Prompts up to 100K tokens cost $0.10 per million input and $0.50 per million output, rising to $0.50 and $2.50 beyond that, up to 90 percent below Haiku 4.5. It offers a 1M context window, up to 128K output tokens, and the first adjustable effort setting on a Haiku model, and targets high-volume work such as summarization, compaction, classification and database queries as well as subagent duty alongside Opus 5.5 and Sonnet 5.5

A multimodal MoE model released by MiniMax, with 428B total parameters, approximately 23B activated, a 1M context, focused on coding and agent workflows, priced at $0.15 per million input tokens

Inception Labs' diffusion-based large language model, claimed to achieve over 1000 tokens/s and 1M context, targeting low-latency, high-throughput scenarios

Alibaba Qwen personal agent that builds and deploys apps from plain chat

Reflection's first open-weight model (Oct 16, 2026): 501B MoE with 23B active, 1M-token context, text-only, Apache 2.0 weights in BF16/FP8/NVFP4, built for coding and agentic work; open-source Mirror CLI and 40+ day-0 partners, founded by ex-DeepMind researchers

OpenAI's AI assistant. On 2026/9/17 it launched ChatGPT for Word, which turns rough notes into a first draft, tightens paragraphs, proofreads, suggests edits and flags formatting problems inside the document, completing the Office trio. On 2026/9/16 it launched Sponsored Agents for ChatGPT Ads, letting users talk to company-sponsored agents after clicking an ad, in testing with select US advertisers, and integrated HubSpot and Shopify. On 2026/7/9 it fully released GPT-5.6 in three tiers — Sol, Terra and Luna — plus the ChatGPT Work agent, and folded Codex into a unified client. On 2026/6/8 OpenAI confidentially filed its S-1 to kick off its IPO.

Anthropic's AI assistant. On 2026/9/16 Claude Cowork and chat merged into a single Claude, with Cowork and Design capabilities available in any conversation, alongside the launch of Claude Docs and Slides and export of presentations to PowerPoint or PDF. It released Claude Sonnet 5 on 2026/6/30 ("Near-Opus thinking, at Sonnet speed"), now the default model for Free and Pro; the flagship Opus 4.8 shipped on 2026/5/28 with a $65 billion Series H at a $96.5 billion valuation. On 2026/6/9 it skipped a generation to launch Claude Fable 5, its first publicly available Mythos-class model, plus Mythos 5, a restricted edition for scientific research.

Multimodal AI model by Google, deeply integrated with Google ecosystem

AI chat assistant by ByteDance, excellent Chinese language experience

Smart assistant by Moonshot AI, supports ultra-long context, excels at document analysis

AI assistant by DeepSeek, outstanding reasoning and coding abilities

AI-powered search engine providing real-time answers with citations

Zhipu AI (Z.ai) self-developed LLM family rooted in Tsinghua tech, known for open weights and strong Chinese-language capability; latest flagship GLM-5.3 targets code and agent tasks with open-sourced weights

ByteDance overseas AI assistant (Doubak international), topped app charts in multiple countries, supporting text/voice/image + AI drawing + agent creation

Meta multimodal model for smart glasses and wearables, powering Meta Glasses 2026

SenseTime daily-refresh model (RiRiXin) for enterprise multimodal AI applications

Large language model by Baidu with powerful Chinese capabilities

AI assistant by Alibaba supporting multimodal interaction

Chat assistant by Zhipu AI, based on GLM large model

AI assistant by Tencent, integrated with WeChat ecosystem and search

Google AI image generation model, high-quality image creation with strong text rendering

Advanced AI image generation model with high-quality output and creative control

OpenAI image model family: GPT-Image-2 launched Apr 21, 2026 (99% Chinese text accuracy, No.1 on LM Arena); upgraded to ChatGPT Images 2.5 on Sep 8, 2026 with up to 50% lower latency, sharper precise editing and stronger multi-turn consistency

Top AI image upscaler and enhancer, stunning detail reconstruction

Global top AI asset and image gen platform with massive design resources

AI photo editing and BG removal, one-click e-commerce photos

World's largest AI image model community, SD ecosystem core

Google multimodal model supporting text, image, audio, and video input/output

ByteDance image generation model, high-quality text-to-image with strong composition

Tencent Hunyuan professional image model preview released 2026/9/22 with text-to-image, image-to-image and multi-turn editing, up to five reference images and 2K output; Tencent Cloud API prices a 2K image at 0.15 CNY and bills output images only, already live in Yuanbao, ima, WorkBuddy and Miora

Alibaba's open-source 7B image model released September 20, 2026, unifying text-to-image and instruction-based editing in one checkpoint with native RGBA transparency, up to 10 reference images and circle, scribble or mask scoping; it took the top open-source spot in both Arena image leaderboards on September 23

AI tool for content generation and creative work

AI image/video detection platform, API support

Multi-functional AI: chat, image gen, writing, 13M monthly visits

Free AI image generation and design platform, generous free tier

AI art community with multiple models for online image generation

AI image editing suite by Stability AI: cutout, BG remove, repaint

AI visual creation tool by Meitu, text/image-to-image, great Chinese support

Professional AI image/video enhancer: denoise, upscale, sharpen

AI art community with multi-model creation and sharing

All-in-one AI design: cutout, AI image gen, editing, leading Chinese tool

Leading Chinese AI image platform and model community, SD core

All-in-one AI art and video creation by Alibaba, popular with designers

AI art model community and online gen platform, mainstream Chinese SD

Krea 3 image editing model: box-select edits, native 2K, Fast and Ultra

Professional camera app with AI features, advanced controls for mobile photography

AI image generation model with improved quality and prompt adherence

AI agent for geographic and environmental data analysis

AI motion generation model for creating realistic character animations

Alibaba's text-to-image model, leading in China for text-to-image generation, supports ultra-long text understanding and fine-grained rendering, API price 0.18 yuan per image

AI image generation tool for e-commerce scenarios, offering AI product images, AI model try-on, background replacement, and poster generation capabilities, with credit-based pricing.

AI solution provider for the apparel industry, offering AI product photography, clothing recognition, and batch product image generation based on its proprietary RIDGE model, serving over 3,000 apparel brands.

Google's image creation and editing tool built on Nano Banana: object-level localized edits, in-image text editing and translation, real-time collaboration and multi-option generations, at pics.new and inside Docs and Slides for Google AI Pro/Ultra and Workspace business users

Bria 8B-parameter open-weight image model, ranked first on the AA image editing leaderboard, covering multiple types of editing actions and text-to-image

AI vector and illustration generator for logos, icons, illustrations

AI one-click background removal, fast and precise, design essential

One-click AI image and audio detection, clean interface

AI image detection, deep analysis, multi-model identification

Google text-to-image diffusion model, high-fidelity image generation with strong prompt understanding

Top AI art tool, generating high-quality artistic images from text descriptions

Image generation model by OpenAI, integrated in ChatGPT

Open-source AI image generation model, self-hostable with rich community

AI image and video generation tool by ByteDance

Open-source image gen model by Black Forest Labs, rivaling closed-source quality

Open-source node-based image generation interface, powerful workflow editor for Stable Diffusion and other models

SenseTime daily-refresh model (RiRiXin) for enterprise multimodal AI applications

Professional AI image generation platform for game and design assets

Alibaba video and image model, Wan3.0: 30s clips, docs, editing

AI image generator strong on text, now text-to-SVG with Vector 1.0

Generative AI by Adobe, deeply integrated into Photoshop and more

Google multimodal model supporting text, image, audio, and video input/output

AI tool for creative content generation

AI video clip and editing tool for creating short-form content

A dual real-time video model released by Shengshu Technology on 2026/9/15, just 69 days after Vidu S1, and made fully available at launch. S2-Avatar raises real-time digital character output from 540P to 720P at 25-42 FPS, and lets users feed in reference images of objects, clothing or scenes at any point during generation while keeping the character holding the object. S2-Editing applies real-time style transfer, virtual try-on, character swaps and background swaps to an incoming video stream. The company is also exploring spatial video for VR headsets.

First preview build of Shengshu Technology next-generation flagship video generation model, released on 2026/10/8 and fully rolled out at launch. Upgrades focus on character performance, camera language and complex visual effects; accepts up to 15 reference images and 3 reference audio tracks, with output covering 2K and 4K, launching at 0.09 yuan per second.

AI video generation product by MiniMax, outstanding multimodal capabilities

Video gen model by Google DeepMind, physical realism + native audio

AI video platform: text-to-video, AI avatars, 11.7M monthly visits

AI long-to-short video tool, one-click clip for social platforms

ByteDance AI video generation model, high-quality text-to-video with motion control

ByteDance video generation model 2.5, 30s single-pass + 50 references + 3D preview + built-in editing, paired with Doubao 2.1 Pro

AI tool for content generation and creative work

Google DeepMind's world model announced 2026/10/28, generating visuals together with ambient audio, music and character dialogue, with keyboard control and real-time natural language edits while Gemini Omni anchors appearance and Gemini 3.8 Flash keeps causality, currently in research preview

AI image/video detection platform, API support

Professional AI image/video enhancer: denoise, upscale, sharpen

All-in-one AI art and video creation by Alibaba, popular with designers

Krea 3 image editing model: box-select edits, native 2K, Fast and Ultra

Enterprise AI digital human video, 160+ languages, 230+ AI avatars

Chinese AI video gen by Shengshu Tech, 1080P HD video support

AI short video generator with rich effects for social media

AI online video editor: auto subtitles, voice translation, avatars

AI video editor by Wondershare, powerful desktop with rich effects

AI text-to-video and voiceover, 80+ languages, 2000+ AI voices

AI text-to-video from scripts/articles/PPTs, 2M monthly visits

AI digital human video platform, leading Chinese digital human product

Leading AI creation: up to 50-min video gen, breakthrough long-video

AI video gen by Tencent, Hunyuan model, multiple styles

AI video creation by Bilibili, deep Bilibili ecosystem integration

Hybrid video diffusion Transformer, supports 720p video generation on a single GPU, utilizing hybrid linear attention and attention residual techniques

MiniMax H3 is a general-purpose full-modal generation system that can uniformly understand multimodal contexts composed of text, images, video, and audio, generating videos with up to 2K resolution, up to 15 seconds in length, and native stereo audio. It supports multiple aspect ratios, 24 FPS output, and stably supports 11 languages. Two versions are available: H3-Base-FL2VA (first-and-last-frame mode) and H3-Base-Ref2VA (full-modal reference mode).

MiniMax's multimodal creative Agent workbench based on the H3 model, automatically decomposes tasks to complete material processing, video generation, and editing delivery, with a built-in 3D director's console

JD.com's open-source real-time streaming video editing system, autoregressive diffusion architecture, 30 frames per second at 720P, supports natural language frame-by-frame editing of live streams

World Labs, led by Fei-Fei Li, released the world's first multimodal world model on 2026/9/1: using a spatial context mechanism to unify text, images, videos, camera poses, and 3D depth anchored in three-dimensional space, enabling pixel-level camera-controlled video generation (up to 1 minute at 1440p), explicit 3D reconstruction from sparse photos (point clouds/3D Gaussian splatting), spatiotemporal simulation, and robot Real-to-Sim; currently open only to early access for partners

The image-to-video entry in HiDream O1 series, which debuted at No.6 on the Artificial Analysis image-to-video with audio leaderboard on October 9, 2026, in the same tier as MiniMax H3, Vidu Q4 Preview and Gemini Omni Flash.

An AI-native creative suite covering video generation, camera control, ad production and AI influencers, bringing together Seedance, Veo, Kling, Sora, Nano Banana and more across Cinema Studio, Ads Studio, Katana and Supercomputer

Viggle 3D native world model that models in three-dimensional space and generates real-time extensible 3D scenes and characters with motion in seconds

Open robot video labeling suite: segment, caption and score training data

AI image and video generation tool by ByteDance

AI video generation and editing platform, text/image to video, used in film industry

AI video generation tool with clean interface, text and image to video

AI video generation tool by Kuaishou, outstanding quality and dynamics

Text-to-video model by OpenAI, up to one-minute high-quality videos

Video editing tool by ByteDance with rich built-in AI features

Google video generation model, high-quality text-to-video with cinematic control

ByteDance Jimeng AI video generation tool with multi-modal input

Alibaba video and image model, Wan3.0: 30s clips, docs, editing

AI digital human video platform, multilingual talking head videos

AI digital human platform, upload photo to generate talking videos

Video gen tool by Luma AI, realistic videos from text and images

AI audio/video editor, edit media by editing text

Bilingual translation extension supporting 20+ engines, highly acclaimed

Chinese AI all-in-one: search, writing, PPT and more

AI copywriting + SEO tool for content marketing optimization

Chinese AI writing tool for official, academic and business genres

Browser AI assistant for writing, translating, summarizing anytime

AI rewriting and polishing tool for clarity and tone

AI writing assistant for content, research, speeches and more

AI fiction and content creation, popular for web novel writing

Free AI writing by iFlytek, multiple genres, great Chinese

AI writing platform by Xinhua News Agency, authoritative for government

AI marketing copy tool for headlines, landing pages, ad copy

OpenAI's AI assistant. On 2026/9/17 it launched ChatGPT for Word, which turns rough notes into a first draft, tightens paragraphs, proofreads, suggests edits and flags formatting problems inside the document, completing the Office trio. On 2026/9/16 it launched Sponsored Agents for ChatGPT Ads, letting users talk to company-sponsored agents after clicking an ad, in testing with select US advertisers, and integrated HubSpot and Shopify. On 2026/7/9 it fully released GPT-5.6 in three tiers — Sol, Terra and Luna — plus the ChatGPT Work agent, and folded Codex into a unified client. On 2026/6/8 OpenAI confidentially filed its S-1 to kick off its IPO.

AI chat assistant by ByteDance, excellent Chinese language experience

Enterprise AI writing tool for marketing copy, blogs, social media

AI writing assistant built into Notion, seamless document workflow

AI English writing assistant with grammar correction and tone adjustment

Deep learning translation tool, widely acclaimed for quality

NetEase translation tool supporting multiple languages, document and voice translation

AI copywriting tool for ad copy, emails, social media posts

Chinese AI writing assistant for rewriting, continuation, correction, translation

AI paraphrasing and summarization tool for better text expression

AI English writing assistant by ByteDance for grammar and polishing

iFlytek Spark chat and voice models, X2.6 with a 1M context window

OpenAI's sixth-generation flagship model released 2026/9/3 (PT): billed as the world's most intelligent and aligned model, built for long-horizon agentic work and computer use; first model to reach the Critical cybersecurity threshold; 99.9% ARC-AGI-3, 97.6% FrontierMath Tier 4, 100% ExploitBench; API $10/$50

ByteDance flagship released 2026/6/23, rivaling GPT-5.5/Opus 4.7; 10+ hour autonomous engineering, 2-hour video understanding, 80% lower cost than Opus

Anthropic's first Claude 5.5 model, released 2026/9/22: matches Claude Fable 5.1 on most work; $4 per million input tokens and $20 per million output, cache reads down 60%, total operating cost about 40% below Opus 5 and output more than 30% faster

OpenAI light-tier GPT-6 model released 2026/9/22 (PT) for high-volume jobs with a clear goal such as summarization, extraction and short answers; API priced at $0.10 per million input tokens and $0.50 per million output, half of GPT-5.6 Luna, available to Free and Go users in the desktop app

Open-source VS Code AI coding plugin (formerly Roo Cline), multi-model + MCP + multi-mode, free alternative to Cursor

ByteDance full-lifecycle AI development platform (formerly TRAE SOLO, renamed TRAE Work on 2026/6/9; TraeWork and TraeCode merged into TRAE on 2026/10/9), switching seamlessly between Agent mode and IDE mode across desktop, web and mobile, with multi-project parallelism, custom agents and Design mode; free tier available, paid tiers raise cloud concurrency and priority

Moonshot coding model released 2026/6/12, 1.1T MoE/32B active/256K context/native INT4; 3 coding benchmarks beat K2.6, 30% less token usage

Zhipu AI coding tool released 2026/6/13, fully switched to self-developed ZCode Agent core, optimized for GLM-5.2 long-range reasoning

Xiaomi MiMo team open-source terminal coding agent released 2026/6/13, based on OpenCode; 1M context, topped GitHub Trending on day one

Microsoft enterprise agent platform launched 2026/6/17, rivaling Claude Cowork; 30-40% lower cost, native M365 integration

Tencent Hunyuan model with strong coding and reasoning, multimodal capabilities

Xiaomi MiMo agent tool for automated task execution

Vercel AI development assistant for building and deploying web apps

AI agent platform for enterprise automation and workflow orchestration

AI software engineer by Cognition AI, autonomous development tasks

Terminal AI coding agent by Anthropic. On 2026/9/17 Projects was redesigned from folders into hosted multi-threaded conversational projects: the user sets a goal, Claude splits the task, schedules threads in parallel, reviews the output and consolidates it. On 2026/6/30 Claude Sonnet 5 was added as the high-frequency execution model with Opus 4.8 handling planning; on 2026/5/28 Dynamic Workflows shipped alongside Opus 4.8, dispatching dozens to hundreds of sub-agents that work in parallel and review each other; over 4 million weekly active users

Free AI coding by Alibaba, multi-IDE, leading Chinese coding assistant

Google command-line AI coding tool powered by Gemini, for terminal-based development

Cursor major UI/UX upgrade April 2026, new Agents Window + multi-agent parallel + Design Mode; mobile app launched 6/30

Product Hunt weekly winner, auto-generates websites from Google Maps reviews, for local business SEO

Rust-based smart terminal with AI command generation/explanation/workflow automation, developer productivity tool

Open-source AI workflow automation, tech team favorite, self-hostable

2026's breakout open-source AI agent, 30+ platforms, autonomous tasks, GitHub stars surging

Autonomous AI agent with persistent memory and self-evolving skills, gets smarter with use

xAI's first terminal coding agent, which added cross-session memory on 2026/9/16. After each conversation turn, it writes project conventions, decisions and facts into markdown notes in the background, scoped by project and globally; /memory offers read-only browsing, /dream organizes them into topic files, and instructions in the current conversation take precedence over the notes. Built on the Grok 4 series, it uses parallel sub-agent mechanisms for a 3x efficiency gain, supports headless mode for automated deployment, and takes aim at Claude Code.

Cursor self-developed coding model 2.5, based on Kimi K2.5 (1.04T MoE) + in-house RL, $0.10/task cheapest frontier coding model

GitHub agent coding environment GA at 2026/6/2 Build; Fleet mode (autonomous CLI) + Autopilot mode, Jira/Datadog/ServiceNow integrations

Microsoft self-developed coding model, 2026/6 Build, MoE + Maia accelerator, strong on Rust/Haskell; replaces GPT-4 Turbo as Copilot default from 2026/8

Anthropic feature launched with Opus 4.8 on 2026/5/28: one workflow command spawns dozens to hundreds of sub-agents, HN 198 points

Cognition (Devin parent) released 2026/6/8, upgrading AI engineer to repo-level collaborator: PR review, refactoring, cross-repo, root-cause analysis

Moonshot AI work platform with document analysis, research, and task automation

Moonshot K3 post-train: 2.8T params, AA index 55, 2 dollars per M input tokens

AI UX design assistant for creating user flows and wireframes

Alibaba Qwen personal agent that builds and deploys apps from plain chat

AI coding assistant by AWS, deep cloud service integration

Developer AI search for coding Q&A and debugging

AI website generator, describe needs to auto-build complete sites

AI IDE by Amazon, deep AWS integration, unified coding and cloud dev

Free AI coding by SenseTime, multi-language, excellent completion

AI coding by Google, Firebase integration, full-stack powerhouse

Israeli AI full-stack app generation tool, conversational building of complete web apps with database, one-click deploy

Governance platform for coding agents: Verifiers write team standards as JSON rules in your repo, scored pass or fail by an LLM judge on every change; plus a skill registry and automated loops for code review and skill optimization, working with Claude Code, Cursor, Codex and any agent

Open-source local AI agent framework, 31 built-in tools, permission hardening, SQLite memory, token budget, CLI + Telegram

Local AI Agent tool, Chinese open-source Manus alternative, free

AI Agentic coding by Alibaba, agent-driven next-gen programming

Chrome DevTools team official open-source, MCP protocol lets AI coding agents natively use browser dev tools for debugging

100% local pre-indexed code knowledge graph, optimizing token usage and privacy for Claude Code/Cursor/Codex, open source

CLI and TypeScript SDK released by OpenAI for finding, verifying, and fixing security vulnerabilities in code, supporting repository scanning, change review, issue tracking, and CI security checks.

Open-source strongest coding and collaboration model, 2.4T parameters, first open-sourcing Qwen-Max level weights

Alibaba Tongyi Qianwen, open-sourced late at night on 2026/8/26, is a pioneer preview model based on the next-generation Qwen4 architecture: multimodal MoE, with a total main model parameter of 125B, activating only 6B per token, plus a 51B N-gram vocabulary layer, natively supporting 256K context expandable to 1M; training cost is nearly 90% lower than Qwen3.7-Plus, with weights open-sourced on Hugging Face and ModelScope

An open-source coding agent with self-improvement capabilities, built on recursive language models (RLM) and a continuous framework, supporting CRUD operations on prompts, skills, memory, and sub-agents.

Mojo 1.0 officially released, providing a stable, production-ready foundation for system programming languages in the AI ecosystem.

Agent-native code hosting platform launched by Cursor, offering repositories, PRs, code browsing, and two-way GitHub sync. Cloud Agents can directly clone branches and commit code.

Slack's programming Agent channel, where teams collaborate with Agents like Claude, Devin, and Copilot in the same channel, supporting code diff review and real-time HTML preview

DeepSeek's open-source agent runtime framework, MIT license, microkernel architecture where everything is a plugin, providing four operating modes: standard, code, minimalist, and creative

Alibaba's open-source 27.8B dense multimodal model, Apache 2.0 license, 262K context, Terminal-Bench 73.0, runs on consumer-grade GPUs

Vercel's open-source lightweight native coding Agent, written in Zig, single binary only ~6MB, cold start 10 microseconds, model-agnostic and embeddable in browsers

Open-source coding Agent memory layer released by Hugging Face on 2026/9/3: local Lance dataset storage + optional HF dataset cloud sync, compatible with Claude Code/Codex/pi/Hermes, etc., hybrid retrieval (vector + BM25 + reranking), cost approximately 1/4 to 1/8 of handwritten handoff solutions

OpenAI's software engineering agent that can execute coding tasks in an independent cloud workspace, supporting reusable cloud environments, cross-device continuation, code review, and parallel task management.

An agent-native software development platform for engineering teams. Its core agent, Droid, can run across desktop applications, command line, web, IDE, and CI/CD. On September 30, 2026, Automations will officially open to all users. After describing a workflow in natural language, Droid can execute it automatically based on schedules, Slack messages, GitHub events, or webhooks, and supports choosing your own model and your own machine. It offers four deployment forms: public cloud, hybrid, on-premises, and air-gapped.

Anthropic launched this opt-in vulnerability scanner for open source on October 8, 2026, inspired by Google OSS-Fuzz and powered by its strongest models including Claude Mythos, at no cost to projects. Core maintainers enroll by opening a pull request against a designated repository, and eligibility follows criteria similar to OSS-Fuzz, judged case by case on impact to infrastructure and user security. Each report carries a self-contained reproducer, an explanation of the flaw, a bisection identifying when the bug was introduced where possible, and a candidate patch. Output is fully model generated with no human review, so reports may be incorrect. In early validation, 85 of 97 critical and high severity findings met the bar for disclosure, with a single false positive

Google introduced this official programmatic source of truth for developer documentation on October 7, 2026, covering Google Cloud, Firebase, Android and more, replacing brittle web scraping with a structured API that serves fresh Markdown formatted documentation. It ships with a gcloud CLI surface, an official agent skill, an API Explorer and multi-language client libraries. Capabilities include semantic and keyword search, intelligent document chunking and grounded question answering. The agent skill connects over standard MCP to Google Antigravity, Claude Code, Cursor, GitHub Copilot and custom frameworks, and uses a multi-step retrieval workflow that inspects documentation chunks before fetching complete pages or answers to keep token use down. The index refreshes frequently to minimise lag between upstream documentation updates and agent access

HumanLayer is a multiplayer coding agent workspace that brings agent sessions, plan artifacts and code diffs into one place for team collaboration. It supports bring-your-own-key for Claude Code, Codex and others, with local and cloud daemons running sessions in parallel. Starter is free for teams of three or fewer and Pro costs $100 per user per month.

JiuwenSwarm is a swarm agent framework from openJiuwen, the open-source community backed by Huawei. Built around a Coordination Engineering paradigm, a Leader agent decomposes tasks, assembles a team of specialists and negotiates dynamically to deliver end-to-end results. Four core components cover the Agent Swarm coordination layer, Swarm Skills for team-level capabilities, the Swarm Skills Hub for sharing, and skill self-evolution. It ships no model of its own and connects to Huawei Cloud MaaS, OpenAI-compatible APIs and local models, supports both HOTS and HITS human-agent modes, and offers desktop, CLI and HarmonyOS PC builds with source on GitHub.

Tencent Cloud open-sourced this team AI collaboration tool on October 10, 2026. Built on Git, it puts a team skills, working standards, tool configuration and project knowledge into a single repository, reviewed, updated and versioned like code, then synced to whatever agent each member uses. It supports 16 agents including WorkBuddy, CodeBuddy, Claude Code, Codex and Cursor, and six Git providers from GitHub and GitLab to GitCode, CNB, TGit and private Git services. MIT licensed, installable as the teamai-cli npm package, and integrated into Tencent Cloud agent management platform ClawPro.

iOS game site built with AI-assisted development. Titles are published on the App Store as one-time purchases with no in-app purchases and no ads, in traditional Chinese and Japanese for the markets in Taiwan, China and Japan.

Open-weight 40B MoE coding model, 2.5B active, 262K context, runs on a laptop

Apache-2.0 OS-level agent sandbox using Seatbelt and bubblewrap isolation

Microsoft multi-agent coding harness, no orchestrator, scales to 1024 agents

33B dense agentic coding model with 512K context under Apache 2.0

AI search for programmers, precise coding Q&A

30B-A4B MoE trained from scratch in Korea, 256K context, Apache 2.0

OpenAI's AI assistant. On 2026/9/17 it launched ChatGPT for Word, which turns rough notes into a first draft, tightens paragraphs, proofreads, suggests edits and flags formatting problems inside the document, completing the Office trio. On 2026/9/16 it launched Sponsored Agents for ChatGPT Ads, letting users talk to company-sponsored agents after clicking an ad, in testing with select US advertisers, and integrated HubSpot and Shopify. On 2026/7/9 it fully released GPT-5.6 in three tiers — Sol, Terra and Luna — plus the ChatGPT Work agent, and folded Codex into a unified client. On 2026/6/8 OpenAI confidentially filed its S-1 to kick off its IPO.

Anthropic's AI assistant. On 2026/9/16 Claude Cowork and chat merged into a single Claude, with Cowork and Design capabilities available in any conversation, alongside the launch of Claude Docs and Slides and export of presentations to PowerPoint or PDF. It released Claude Sonnet 5 on 2026/6/30 ("Near-Opus thinking, at Sonnet speed"), now the default model for Free and Pro; the flagship Opus 4.8 shipped on 2026/5/28 with a $65 billion Series H at a $96.5 billion valuation. On 2026/6/9 it skipped a generation to launch Claude Fable 5, its first publicly available Mythos-class model, plus Mythos 5, a restricted edition for scientific research.

AI assistant by DeepSeek, outstanding reasoning and coding abilities

Zhipu AI (Z.ai) self-developed LLM family rooted in Tsinghua tech, known for open weights and strong Chinese-language capability; latest flagship GLM-5.3 targets code and agent tasks with open-sourced weights

AI programming assistant by GitHub for code completion and generation

AI-native code editor with deeply integrated AI coding workflow

AI programming assistant by Tencent Cloud, multi-IDE and language support

AI UI generation tool by Vercel, React components via conversation

AI full-stack tool by StackBlitz, build apps via conversation

AI full-stack app generator, create web apps through conversation

OpenAI open-source terminal coding agent with sandboxed execution, for command-line AI programming

Open-source AI pair programming tool in the terminal with smart Git integration and broad LLM support (Apache-2.0)

Open-source VS Code AI coding agent, autonomous file editing and terminal commands

Open-source AI software engineer agent (formerly OpenDevin), autonomous coding and task execution

Google AI agent framework for building autonomous agents with Gemini, announced at I/O 2026

High-performance code editor written in Rust, millisecond startup, built-in AI assistant + real-time collaboration

AI code review tool, auto-analyzes PRs with improvement suggestions, 10M+ PRs reviewed, supports GitHub/GitLab

Block open-source local AI agent framework, autonomous engineering tasks, supports any LLM and MCP

Open-source visual LLM orchestration tool, drag-and-drop RAG/Agent/chat apps, 35K+ GitHub stars

LangChain visual builder, drag-and-drop LLM apps and Agent flows, 45K+ GitHub stars

Browser automation AI agent, executes form filling/data scraping/system operations on web pages without code

Open-source browser agent framework, 60K+ GitHub stars, giving LLMs web navigation and task execution

AI assistant with file system access, code execution, and multimodal generation capabilities

Tencent AI design agent collaboration platform, open beta since May 18 and iterating to v2.7+: editable text-to-UI, Figma drag-and-drop import, MCP that delivers designs straight into CodeBuddy/Cursor for code; Windows client available, 1,000 free credits on signup

Desktop app for running local LLMs, download and chat with open-source models offline

Chat assistant by Zhipu AI, based on GLM large model

Online IDE + AI assistant, write and deploy code in browser

AI code completion tool with local models, focused on privacy

Free AI code completion tool supporting 70+ languages

Open-source terminal AI coding agent, framework for building coding assistants

AI coding assistant with deep codebase context understanding, enterprise-grade security

Open-source AI code assistant for VS Code and JetBrains, autocomplete + chat with any model

ByteDance AI video generation model, high-quality text-to-video with motion control

Microsoft open-source voice AI model family with TTS and ASR, supporting 60-min long audio recognition and 90-min multi-speaker synthesis

ByteDance Doubao audio model released at 2026/6/23 FORCE, supporting reference generation and long-duration multi-role voice consistency

A speech model series released by StepFun on 2026/9/15, comprising five models — Realtime, ASR, TTS, Gen and Music — launched simultaneously on its open platform. Realtime is natively full-duplex, ranking first worldwide on Conversational Dynamics with a composite score of 98.9% and first in speech reasoning accuracy at 99.7%. ASR ties for first worldwide with a 1.7% non-streaming word error rate. TTS reproduces paralinguistic cues such as laughter, hesitation and stuttering. Gen generates vocals, sound effects, ambient sound and BGM from a single instruction, while Music supports multi-turn composition using ABC notation.

Google DeepMind music generation model producing 44.1kHz stereo audio, with custom lyrics, section tags and tracks up to about three minutes, all SynthID watermarked

Full-duplex speech model trained on conversation audio alone, runs on a MacBook

AI transcription assistant by Alibaba Cloud for meetings and translation

AI meeting notes platform, auto-transcription, 60+ languages

AI research tool by Google, doc Q&A + audio overview, free

FaceBrain open-source multimodal model, vision/audio/text, efficient on-device deployment

Pro AI music detection, 99.42% accuracy, 9 AI music generators

AI text-to-video and voiceover, 80+ languages, 2000+ AI voices

AI text-to-speech tool, multilingual, leading TTS product

AI voice cloning platform, clone voices from minimal audio data

Free online AI TTS tool, 70+ languages, 200+ voices

AI noise cancellation + meeting summary, real-time noise removal

AI TTS platform with ultra-realistic voices, voice cloning and API

AI podcast tool: record/edit/transcribe/publish all-in-one

Free AI music creation by ByteDance, multi-style song generation

AI voiceover tool with lifelike quality, leading Chinese AI voice

Mac local voice dictation tool, recognition is completed entirely on-device, inserts into any application following the cursor, supports 110 languages and file transcription

Speech-to-speech model from xAI released on 2026/10/14, first on the Artificial Analysis Speech-to-Speech Index at 86.9, 97% on Full Duplex Bench, time to first audio under one second, API priced the same as its predecessor

Mistral's speech model family, featuring speech transcription and synthesis, covering more than 50 languages, with a companion Voice Agents API, available for EU hosting or self-hosting

Ant Bailing's open-source speech series, comprising five models for speech synthesis, audio understanding, and recognition, supporting seven languages and two dialects

IBM 2B open speech model for six-language transcription and translation

AI meeting notes for Zoom/Meet/Teams transcription and summary

Xiaomi MiMo text-to-speech model with natural voice synthesis

One-click AI image and audio detection, clean interface

Free AI music detection, Suno/Udio identification, audio analysis

Online AI music detection, upload audio, multi-format support

AI music platform with the V6 model family (v6/v6-wild/v6-mini), natural-language section editing, multi-source mashups, image and video input, and the first licensed-data training

AI music generation tool, multiple styles and high-quality audio

AI voice synthesis and cloning platform, realistic human voices

AI music generation tool by MiniMax, launched Music 2.6 model in April 2026 with Cover feature, leading Chinese song quality

AI audio/video transcription tool with 99.8% accuracy, 98+ languages, free unlimited duration, no registration required

iFlytek Spark chat and voice models, X2.6 with a 1M context window

AI voiceover tool with 120+ realistic AI voices

AI music composition assistant for film, game and scene scoring

AI audio/video editor, edit media by editing text

AI music creation platform by NetEase, Chinese song generation

AI meeting notes tool, real-time transcription and summary

Microsoft AI assistant with Bing search and M365 ecosystem

Xiaomi MiMo agent tool for automated task execution

Moonshot AI work platform with document analysis, research, and task automation

Mistral AI optical character recognition model, extracting text from documents and images

AI text-to-diagram tool, one-click flowcharts and mind maps

Chinese AI all-in-one: search, writing, PPT and more

AI transcription assistant by Alibaba Cloud for meetings and translation

AI meeting notes platform, auto-transcription, 60+ languages

Microsoft enterprise AI agent management platform, unified governance of identity, security, and lifecycle for all AI agents

AI automation connecting 7000+ apps, natural language workflows

Free all-in-one AI office agent by ByteDance, custom agents/workflows

Tencent AI knowledge work partner, fully open since 5/25, Copilot Q&A + knowledge plaza + skill marketplace

Luoji Siwei AI note tool upgraded from Get Notes on 5/29, with AI auto-linking knowledge and discovering blind spots

Kingsoft AI note product released 2026/6/7, AI throughout record-organize-reuse flow, two-way sync with WPS Docs/PPT

Data analysis agent inside ChatGPT Work, invoked as @Data in the plugin directory; connects business data sources and turns plain-language questions into shareable dashboards

Xiaomi desktop client for the MiMo family, out of invite-only testing into general availability on September 22, 2026 with membership plans, shipping the three V2.6 models alongside support for your own API key, task decomposition, tool calling, parallel jobs and artifact management

Moonshot AI renamed Kimi WebBridge to the Kimi browser extension on September 22, 2026, adding a sidebar to navigate, fill forms and complete multi-step web tasks via chat, with repeated actions recordable as replayable skills; it drives the browser locally over the Chrome DevTools Protocol so cookies and login sessions never leave the machine, and works alongside Claude Code, Cursor, Codex and other agents

AI writing platform by Xinhua News Agency, authoritative for government

33B dense agentic coding model with 512K context under Apache 2.0

AI meeting notes for Zoom/Meet/Teams transcription and summary

AI project management and team collaboration with task automation

Chinese AI PPT auto-generator, one-click presentations from topics

AI collaborative whiteboard for brainstorming and project planning

AI database/spreadsheet with custom workflows and automation

AI in Slack: smart summary, search, workflow automation

Enterprise AI search and knowledge management, unified app search

AI PPT generator, Google Slides plugin, text to presentation

Document assistant by Baidu: processing, creation, analysis

AI doc generator by iFlytek, one-click PPT and Word creation

AI data analyst by SenseTime, industry-leading spreadsheet/data

Visual automation platform (ex-Integromat), drag-and-drop workflows

Open-source automation platform, no-code 100+ app connections, self-hostable

Baidu AI assistant with PC-mobile handoff, desktop built-in browser, cross-device task execution; smart routing cuts task time 20%

AI collaboration super app originally launched by Instabase, bringing multiple models like Claude, GPT, Gemini, and Grok into the same thread, where conversations can directly become documents, presentations, spreadsheets, and interactive applications

An intelligent document scanning app under IntSig Information, offering AI capabilities such as OCR text recognition, ID card scanning, wrinkle and shadow removal, and table recognition.

A cloud video conferencing product under Tencent, offering AI capabilities such as AI assistant smart minutes, Yuanbao minutes, and AI hosting, covering the entire process before and after meetings.

A free online survey platform under Tencent, powered by the Hunyuan large model, offering end-to-end capabilities including AI-generated questionnaires, intelligent data cleaning, AI reports, and insight summarization.

The intelligent customer service product under Tencent Qidian, providing omnichannel online customer service and customer service robots based on large models, officially claiming to save 68% of labor costs.

LlamaIndex doc parsing: PDFs and tables into Markdown, powered by OCRia-1

SoMark's open-source 1.3B document parser released in late October 2026, outputting Markdown or JSON with layout coordinates and reading order, ranked first on olmOCR-Bench at 87.84, OmniDocBench at 97.18 and DocParsingBench at 87.76, with a replaceable layout module

OpenBMB's Apache 2.0 document recognition suite released 2026/10/27, pairing a 1.3B parser with the 2B Jev quality model that scores text, formulas and tables from 0 to 1, with 93.68 on Real5-OmniDocBench

Baidu's open-source long-document parser released 2026/10/27, whose SR-SWA sparse reference attention compresses cross-page visual links into a fixed reference set and drops peak single-sequence KV memory from 47GB to 1.77GB, parsing an arbitrary number of pages in one pass

Open-source macOS assistant from a YC Fall 2026 team that runs Gemma 4 E4B locally overnight to read mail, messages and files into a Markdown knowledge base, works on an 8GB Apple silicon Mac, AGPL-3.0

AI instant prototyping, PM and designer collab tool, leading Chinese

Proactive agent on Opus 5.5: briefs, plus tasks run in its own cloud

AI writing assistant built into Notion, seamless document workflow

Open-source AI software engineer agent (formerly OpenDevin), autonomous coding and task execution

AI presentation generator, one-click beautiful PPT creation

AI data analysis tool, upload data for auto analysis and visualization

AI assistant with file system access, code execution, and multimodal generation capabilities

Alibaba DingTalk AI assistant for enterprise communication, task automation, and meeting summaries

Tencent PC Manager WeChat remote AI assistant, control computer via WeChat for files, screenshots, and apps

AI business automation platform, connecting 3000+ apps including email/calendar/CRM for automated workflows

OpenAI team-level workspace agents released 2026/4/23, Codex-powered, 24/7 cloud automation of complex workflows

AI features in Figma, auto-layout and design generation

AI tool for chatting with PDF documents, asking questions and extracting information from any PDF

AI document analysis tool, chat with files, instant answers and summaries from long documents

Online IDE + AI assistant, write and deploy code in browser

AI presentation design tool with automatic layout beautification

AI spreadsheet tool, operate Excel with natural language

AI narrative presentation tool, auto-generating text and image slides

AI assistant in Feishu for docs, sheets, meetings and more

AI meeting notes tool, real-time transcription and summary

AI office assistant in WPS for documents, sheets, PPT

Top AI image upscaler and enhancer, stunning detail reconstruction

Global top AI asset and image gen platform with massive design resources

Tencent Hunyuan model with strong coding and reasoning, multimodal capabilities