Agents

Agent Gallery

60 curated AI agents. Source-based assessments, editorial picks, and AI estimates are labeled separately; inspect the evidence before choosing.

Inspect assessment sources, versions, and scenarios →
60 results
94
Featured
Cursor

Catalog description · not a benchmark

The AI IDE with the strongest coding capabilities, putting Claude Opus 4.8 + GPT-5.5 into your editor

Claude Opus 4.8Claude Sonnet 4.6Version checked Nov 17
CodingL3免费
93
Featured
Claude

Catalog description · not a benchmark

Anthropic's AI assistant for drafting, editing and analyzing text.

Claude Opus 4.8Claude Sonnet 4.6Version pending confirmation
WritingL2$20/月
91
Featured
DeepSeek Chat

Catalog description · not a benchmark

The ceiling of cost-effectiveness: #1 in Chinese understanding, native million-token context, API cost less than 1/10 of Claude's

DeepSeek-V4DeepSeek-V4 ProVersion pending confirmation
WritingL2免费网页
89
Midjourney

Catalog description · not a benchmark

Provides image generation from text prompts and image references on the Create page

MidjourneyVersion pending confirmation
CreativeL1需付费订阅;价格与额度以官方方案为准
88
ElevenLabs

Catalog description · not a benchmark

The most natural AI speech synthesis and cloning, the de facto standard for dubbing/audiobooks

Eleven v3Music v2Version checked Nov 17
VoiceL1免费
88
Figure 03

Catalog description · not a benchmark

BotQ factory rolls 1 unit off the line per hour, already commercially billed at $25/robot-hour on BMW's production line

Helix 视觉-语言-动作模型Version checked Nov 17
RoboticsL4企业部署
88
GitHub Copilot

Catalog description · not a benchmark

The earliest and most widely adopted AI coding assistant, deeply integrated with GitHub workflows

GPT-5.5Claude Opus 4.8Version checked Nov 17
CodingL2免费
88
Featured
Kimi

Catalog description · not a benchmark

256K context + Agent Swarm, a new domestic benchmark for long-document understanding and Chinese reasoning

Kimi K2.6Kimi K2.7 CodeVersion pending confirmation
ResearchL4免费
88
Featured
OKCodex Topic Agent

Catalog description · not a benchmark

A topic-selection Agent tailored for Chinese WeChat Official Accounts, generating 5 well-sourced topics in 30 seconds

DeepSeek-V4Qwen3.7-MaxVersion pending confirmation
WritingL2免费试用
88
v0 by Vercel

Catalog description · not a benchmark

Natural language generates runnable React UI, the fastest path to front-end prototypes

Claude Opus 4.8GPT-5.5Version checked Nov 17
CodingL2免费
87
Dify

Catalog description · not a benchmark

Open-source LLM application development platform, 100K+ GitHub Stars, self-hosted with zero lock-in

DeepSeek-V4Qwen3.7Version checked Nov 17
CodingL3开源免费
87
Gemini 3 Pro

Catalog description · not a benchmark

Million-token context + Deep Think: the top choice for large file and long video processing

Gemini 3 ProVersion checked Nov 17
ResearchL2免费
87
NVIDIA Jetson Thor

Catalog description · not a benchmark

The de facto standard for robot brains: an edge computing platform for real-time multi-sensor physical AI

Blackwell GPUIsaac 平台Version pending confirmation
AI HardwareL1开发套件
86
Qwen3.7-Max

Catalog description · not a benchmark

Alibaba Cloud Tongyi Qianwen Flagship Edition: Million-Token Context + Ultra-Long-Horizon Agent Execution

Qwen3.7-MaxQwen3.7-PlusVersion checked Nov 17
ResearchL3免费体验
86
Runway

Catalog description · not a benchmark

The benchmark for AI video generation, the Gen series lowers the barrier to short film creation

Gen-4.5Gen-4Version checked Nov 17
CreativeL1免费额度
86
Featured
Unitree G1

Catalog description · not a benchmark

The Chinese humanoid robot that appeared on the Spring Festival Gala is now listed on Amazon ($17,990), leading globally in mass production shipments

具身智能控制栈Version checked Nov 17
RoboticsL4$17,990 起
86
Windsurf

Catalog description · not a benchmark

Agentic IDE: Cascade-centric automated coding, Cursor's main rival

Claude Opus 4.8GPT-5.5Version pending confirmation
CodingL3免费
85
n8n

Catalog description · not a benchmark

Open-source workflow automation + AI Agent nodes, a self-hostable Zapier alternative

可接入任意 LLMVersion checked Nov 17
OperationsL3自托管免费
85
Perplexity AI

Catalog description · not a benchmark

Provides web search and source citations for checking public information

Version pending confirmation
ResearchL1免费基础搜索
85
Featured
Ray-Ban Meta

Catalog description · not a benchmark

Best-selling AI hardware: the Llama 4-powered "Hey Meta" portable assistant, aiming for ten million units in the second half of 2026

Llama 4Meta AIVersion pending confirmation
AI HardwareL1$299 起
84
Gamma

Catalog description · not a benchmark

Generate presentations/webpages/documents from one sentence, the fastest AI layout tool for making slides

GPT-5.5Version pending confirmation
ProductivityL2免费
84
Manus

Catalog description · not a benchmark

Asynchronous full-stack Agent: Let AI truly complete tasks for you

Claude Opus 4.8DeepSeek-V4Version checked Nov 17
BrowserL4邀请制
84
Plaud Note Pro

Catalog description · not a benchmark

Forbes 2026 Best AI Wearable of the Year: Portable recording + AI notes for meetings/consultations/interviews

GPT-5.5Claude Opus 4.8Version checked Nov 17
AI HardwareL1硬件 + 订阅
84
Qwen-Audio-TTS

Catalog description · not a benchmark

阿里云百炼's Chinese-first speech synthesis API, covering real-time, non-real-time, and voice cloning

qwen-audio-3.0-tts-plusqwen-audio-3.0-tts-flashVersion checked Nov 17
VoiceL2API 按量(以百炼实时价目表为准)
84
Suno

Catalog description · not a benchmark

Enter lyrics and style to generate a song in one click—the most breakout product in AI music generation

Suno v5.5Version checked Nov 17
CreativeL1免费
83
Notion AI

Catalog description · not a benchmark

Write AI into your knowledge base: a workflow assistant that integrates documents/meetings/databases

GPT-5.5Claude Opus 4.8Version checked Nov 17
ProductivityL2$10/成员·月(附加项)
82
Featured
扣子(Coze 国内版)

Catalog description · not a benchmark

Agent building platform by ByteDance, 80% revenue share, quickest to get started

豆包MiniMaxVersion checked Nov 17
ProductivityL3免费
82
AI estimate
LiveKit Agents

Catalog description · not a benchmark

A framework for building real-time multimodal AI applications

Version checked Nov 17
VoiceL3开源,云服务付费
82
AI estimate
Macro

Catalog description · not a benchmark

AI-powered browser automation and web operation tool

Version checked Nov 17
BrowserL3免费开源,可能有付费云服务
82
AI estimate
OpenClaw

Catalog description · not a benchmark

An open-source personal AI assistant Agent that runs on your own machine

Version checked Nov 17
ProductivityL4开源免费(自备模型 API Key)
80
1X NEO

Catalog description · not a benchmark

The first humanoid robot to enter the home: pre-orders are open, with delivery to U.S. families in late 2026

1X 端到端神经控制Version checked Nov 17
RoboticsL4预订制
80
Grammarly

Catalog description · not a benchmark

The long-standing standard for English writing error correction and polishing, available at hand across all platforms

自研 NLPGPT-5.5Version checked Nov 17
WritingL1免费
80
Copilot in PowerPoint

Catalog description · not a benchmark

Generate and continue editing presentations within Microsoft 365 from prompts or existing files

Microsoft 365 CopilotVersion checked Nov 17
ProductivityL2需 Microsoft 365
79
Tesla Optimus Gen 3

Catalog description · not a benchmark

Tesla discontinues Model S/X to free up its production line: the first humanoid robot designed for large-scale mass production

FSD 神经网络栈Version pending confirmation
RoboticsL4未发售
79
通义千问 Agent

Catalog description · not a benchmark

阿里云 ecosystem Agent: enterprise-grade tool invocation + domestic compliant deployment

Qwen3.7-MaxQwen3.7-PlusVersion checked Nov 17
OperationsL3免费体验
78
Gemini Notebook

Catalog description · not a benchmark

AI note and research assistant, integrating Gemini capabilities

GeminiVersion checked Nov 17
ResearchL3免费
78
智谱清言 GLM-5

Catalog description · not a benchmark

Chinese-language Agent with a Tsinghua technology background, GLM-5.2 one-million-token context, capabilities benchmarked against Opus 4.8

GLM-5.2GLM-5VVersion checked Nov 17
ResearchL2免费
76
AI estimate
Supermemory

Catalog description · not a benchmark

A self-hostable long-term memory layer for AI applications and agents

Version checked Nov 17
DataL2开源免费自托管
74
Doubao

Catalog description · not a benchmark

ByteDance's AI assistant for writing, web summaries and desktop workflows.

豆包大模型Version checked Nov 17
WritingL1免费
74
YuE

Catalog description · not a benchmark

Open-source music large model for full-song generation, from lyrics to complete songs

YuE2Version checked Nov 17
CreativeL2免费(开源,需自备算力)
72
AI estimate
AI Berkshire

Catalog description · not a benchmark

AI-powered Berkshire investment analysis tool

Version checked Nov 17
DataL2免费开源
72
AI estimate
Airi

Catalog description · not a benchmark

Open-source multimodal AI assistant that can be self-hosted

Version checked Nov 17
MultimodalL2免费开源
72
AI estimate
Bond

Catalog description · not a benchmark

Self-executing AI to-do list

Version pending confirmation
ProductivityL3免费
72
AI estimate
Fundraisly

Catalog description · not a benchmark

AI fundraising agent, automatically finds investors and books meetings

Version pending confirmation
SalesL3免费试用
72
AI estimate
gstack

Catalog description · not a benchmark

AI-powered full-stack application generator

Version checked Nov 17
CodingL3免费开源
72
AI estimate
Mailwarm 2.0

Catalog description · not a benchmark

Email warm-up tool to improve deliverability

Version pending confirmation
OperationsL2免费试用
72
AI estimate
MathCode

Catalog description · not a benchmark

Math coding agent that automatically generates and verifies solution code

Version checked Nov 17
CodingL3开源免费
72
MinerU

Catalog description · not a benchmark

Open-source PDF document parsing and data extraction tool

自定义深度学习模型Version checked Nov 17
DataL2免费开源
72
AI estimate
Page Agent

Catalog description · not a benchmark

Alibaba's open-source web automation AI agent

未知Version checked Nov 17
BrowserL3开源免费
72
AI estimate
SellerClaw

Catalog description · not a benchmark

Cross-channel store AI agent team

Version pending confirmation
OperationsL3免费试用
72
AI estimate
Upstream

Catalog description · not a benchmark

The intelligent inbox designed for humans and AI agents

Version pending confirmation
ProductivityL3免费
68
AI estimate
OpenMontage

Catalog description · not a benchmark

Open-source AI montage video generation tool

Version checked Nov 17
CreativeL2免费开源
68
AI estimate
OpenSpec

Catalog description · not a benchmark

AI-driven technical specification generation tool

Version checked Nov 17
CodingL2开源免费
68
AI estimate
semantica-agi/semantica

Catalog description · not a benchmark

Semantics-driven autonomous agent framework

Version checked Nov 17
CodingL3开源免费
68
AI estimate
VC Boom

Catalog description · not a benchmark

AI-driven fundraising matching platform that optimizes pitch materials and matches investors

Version pending confirmation
SalesL2免费试用
62
minimi

Catalog description · not a benchmark

Claude's peripheral memory assistant

ClaudeVersion pending confirmation
ProductivityL3免费
56
AI estimate
MicroLLM Lab

Catalog description · not a benchmark

Run 7 small models directly in the browser, try LLM with zero installation

Version pending confirmation
BrowserL1免费
Unrecorded
DeepSeek API

Catalog description · not a benchmark

Provides function calling, JSON output and streaming output through a developer model API

deepseek-flashdeepseek-v4-proVersion checked Nov 17
CodingL1API 按用量计费;以官方价格与账户条件为准
Unrecorded
Doc2X Document Parsing API

Catalog description · not a benchmark

Document parsing and OCR; API export scope differs from the web app, results unmeasured

Version checked Nov 17
DataL1API 按用量计费;与网页会员分开,实际价格和配额以官方当前页面为准
Unrecorded
Ollama Local Inference

Catalog description · not a benchmark

Supports offline inference with prepared local models and your own compute

Version checked Nov 17
CodingL1本地软件免费;自备硬件,模型许可及运维成本另核;不含云服务

The Codex Score is out of 100. The catalog separates verifiable reviews from clearly labeled AI estimates. Learn about the methodology →