资讯板块落地页

Cursor's Composer 2 was secretly built on a Chinese AI model — and it exposes a deeper problem with Western open-source AI

Technology ‧ sam.witteveen@venturebeat.com (Sam Witteveen)

Luma AI launches Uni-1, a model that outscores Google and OpenAI while costing up to 30 percent less

Technology ‧ michael.nunez@venturebeat.com (Michael Nuñez)

Nvidia's Nemotron-Cascade 2 wins math and coding gold medals with 3B active parameters — and its post-training recipe is now open-source

Orchestration ‧ sam.witteveen@venturebeat.com (Sam Witteveen)

The three disciplines separating AI agent demos from real-world deployment

Orchestration ‧ taryn.plumb@venturebeat.com (Taryn Plumb)

Show us your agents: VB Transform 2026 is looking for the most innovative agentic AI technologies

Technology

From lab to market: Rose Rock Bridge fast-tracks energy innovation in Tulsa

Technology

Testing autonomous agents (Or: how I learned to stop worrying and embrace chaos)

Orchestration

You thought the generalist was dead — in the 'vibe work' era, they're more important than ever

Technology

$Mistral's Small 4 consolidates reasoning, vision and coding into one model — at a fraction of the inference cost$

Mistral's Small 4 consolidates reasoning, vision and coding into one model — at a fraction of the inference cost

Technology

Scale AI launches Voice Showdown, the first real-world benchmark for voice AI — and the results are humbling for some top models

Data ‧ carl.franzen@venturebeat.com (Carl Franzen)

Three ways AI is learning to understand the physical world

Technology ‧ bendee983@gmail.com (Ben Dickson)

Anthropic just shipped an OpenClaw killer called Claude Code Channels, letting you message it over Telegram and Discord

Orchestration ‧ carl.franzen@venturebeat.com (Carl Franzen)

Meta's rogue AI agent passed every identity check — four gaps in enterprise IAM explain why

Security ‧ louiswcolumbus@gmail.com (Louis Columbus)

Cursor’s new coding model Composer 2 is here: It beats Claude Opus 4.6 but still trails GPT-5.4

Technology ‧ carl.franzen@venturebeat.com (Carl Franzen)

Why enterprises are replacing generic AI with tools that know their users

Infrastructure ‧ taryn.plumb@venturebeat.com (Taryn Plumb)

$Xiaomi stuns with new MiMo-V2-Pro LLM nearing GPT-5.2, Opus 4.6 performance at a fraction of the cost$

Xiaomi stuns with new MiMo-V2-Pro LLM nearing GPT-5.2, Opus 4.6 performance at a fraction of the cost

Technology ‧ carl.franzen@venturebeat.com (Carl Franzen)

New MiniMax M2.7 proprietary AI model is 'self-evolving' and can perform 30-50% of reinforcement learning research workflow

Technology ‧ carl.franzen@venturebeat.com (Carl Franzen)

Enterprise AI agents keep operating from different versions of reality — Microsoft says Fabric IQ is the fix

Data

Nvidia says it can shrink LLM memory 20x without changing model weights

Orchestration ‧ bendee983@gmail.com (Ben Dickson)

The authorization problem that could break enterprise AI

Security

首页 > 科技 > VentureBeat

VentureBeat

Cursor's Composer 2 was secretly built on a Chinese AI model — and it exposes a deeper problem with Western open-source AI

Luma AI launches Uni-1, a model that outscores Google and OpenAI while costing up to 30 percent less

Nvidia's Nemotron-Cascade 2 wins math and coding gold medals with 3B active parameters — and its post-training recipe is now open-source

The three disciplines separating AI agent demos from real-world deployment

Show us your agents: VB Transform 2026 is looking for the most innovative agentic AI technologies

From lab to market: Rose Rock Bridge fast-tracks energy innovation in Tulsa

Testing autonomous agents (Or: how I learned to stop worrying and embrace chaos)

You thought the generalist was dead — in the 'vibe work' era, they're more important than ever

Mistral's Small 4 consolidates reasoning, vision and coding into one model — at a fraction of the inference cost

Scale AI launches Voice Showdown, the first real-world benchmark for voice AI — and the results are humbling for some top models

Three ways AI is learning to understand the physical world

Anthropic just shipped an OpenClaw killer called Claude Code Channels, letting you message it over Telegram and Discord

Meta's rogue AI agent passed every identity check — four gaps in enterprise IAM explain why

Cursor’s new coding model Composer 2 is here: It beats Claude Opus 4.6 but still trails GPT-5.4

Why enterprises are replacing generic AI with tools that know their users

Xiaomi stuns with new MiMo-V2-Pro LLM nearing GPT-5.2, Opus 4.6 performance at a fraction of the cost

New MiniMax M2.7 proprietary AI model is 'self-evolving' and can perform 30-50% of reinforcement learning research workflow

Enterprise AI agents keep operating from different versions of reality — Microsoft says Fabric IQ is the fix

Nvidia says it can shrink LLM memory 20x without changing model weights

The authorization problem that could break enterprise AI

相关分类