中文

AI Agent News

Track major events, funding, model releases and breakthroughs across the AI Agent landscape

AI Agent updates

Latest industry news

Track major events, funding, model releases and breakthroughs across the AI Agent landscape

Key events timeline

2026-01

OpenClaw erupts on GitHub

OpenClaw hits global GitHub Top 10 in 10 days, outpacing the Linux kernel star growth

2025-12

Meta acquires Manus for $2B

Meta acquires Manus AI for $2B, locking in the general-purpose Agent race

2025-04

DeepSeek-V3 open-sourced

The value king, at just 5% of GPT-4 cost

2025-03

Manus goes viral overnight

The world's first general-purpose AI Agent draws unprecedented attention

2025-02

OpenAI Deep Research

OpenAI ships a deep-research Agent that generates professional reports in one click

2025-02

MCP Servers pass 500

The MCP ecosystem erupts — 500+ servers built in 3 months

2025-01

DeepSeek-R1 stuns the world

Open-source reasoning model at just 3% of OpenAI cost, reshaping the global AI landscape

2024-11

MCP protocol born

Anthropic releases the Model Context Protocol, the de facto standard for Agent interfaces

2024-10

Claude Computer Use

Anthropic lets AI directly control the computer screen for the first time, opening a new paradigm

2024-09

Replit Agent full-stack automation

Natural language to a shipped product, aimed at non-engineers

2024-08

Cursor ARR passes $100M

The fastest-growing SaaS ever, the new king of AI coding tools

2024-06

Claude 3.5 tops SWE-bench

The strongest coding AI, bug-fixing at a junior engineer level

2024-03

Devin launches

The world's first autonomous AI software engineer, able to complete full coding tasks on its own

Model ReleaseSep 11, 2025

Mistral Releases Large 2: Performance on Par with GPT-4o, Open Source Path Continues

Mistral AI has released Mistral Large 2 (123B parameters), achieving performance comparable to GPT-4o in code generation, multilingual processing, and instruction following, while maintaining the Apache 2.0 open-source license. It supports 32K context, over 80 languages, and is optimized for code. Europe's largest AI unicorn, Mistral, is now valued at €6 billion.

Model ReleaseAug 18, 2025

Baidu ERNIE 5.0 Released: Native Multimodal, Reasoning Ability on Par with GPT-4o

Baidu released ERNIE 5.0, which surpasses GPT-4o across Chinese benchmarks and matches it in multilingual reasoning tests. It adds native video understanding (supporting 2-hour long video analysis) and enhanced coding ability (87.5% on HumanEval). ERNIE 5.0 is integrated into Baidu Search, Baidu Library, and Baidu Netdisk, with 50 million daily active users.

Model ReleaseMay 20, 2025

DeepSeek R2 Launch: Matching GPT-o1 at 1/10 the Cost, Another Milestone for Chinese AI

DeepSeek releases its R2 reasoning model, achieving parity with GPT-o1 on AIME math competition and Codeforces programming benchmarks, yet at only 10% of the training cost. The model is fully open-source and has garnered over 100,000 GitHub stars. The R2 launch has reignited global discussions on AI cost competition, with OpenAI and Anthropic subsequently announcing price cuts.

Model ReleaseMay 17, 2025

Meta Releases Llama 4 Series: Scout and Maverick Models Open-Sourced, Performance on Par with GPT-4o

Meta AI has released the Llama 4 series, including Scout (17B active parameters) and Maverick (17B active parameters, 128-expert MoE). Maverick matches GPT-4o on multiple benchmarks, while Scout focuses on low-latency inference. Both models are fully open-sourced, commercially usable, and already integrated into local inference frameworks like Llama.cpp and Ollama.

Model ReleaseMay 14, 2025

Anthropic Launches Claude 3.7 Sonnet: Code Capabilities Leap Forward, Supports Extended Thinking

Anthropic releases Claude 3.7 Sonnet, achieving a breakthrough score of 62% on SWE-bench (software engineering benchmark), surpassing all existing models. A new "Extended Thinking" mode allows the model to perform deep internal reasoning before answering, improving accuracy on complex math and logic problems by 35%.