SIGNAL
Tracking the global AI frontier — labs · research · agents · policy
Frontier Signal

Labs

Every move from OpenAI, Anthropic, Google DeepMind, Meta, Mistral, xAI, DeepSeek, Qwen, Moonshot and the global frontier labs: model launches, research drops, product news.

Every move from OpenAI, Anthropic, Google DeepMind, Meta, Mistral, xAI, DeepSeek, Qwen, Moonshot and the global frontier labs: model launches, research drops, product news.

Latest in Labs

30 stories
Claude Code v2.1.172
Agents

Claude Code v2.1.172

OpenAI Agents SDK v0.17.5
Agents

OpenAI Agents SDK v0.17.5

Claude Code v2.1.173
Agents

Claude Code v2.1.173

Claude Code v2.1.174
Agents

Claude Code v2.1.174

Claude Code v2.1.175
Agents

Claude Code v2.1.175

Anthropic built a model too risky to release
Briefing

Anthropic built a model too risky to release

and Meta makes an unexpected entry…

That's my designer - Claude
Briefing

That's my designer - Claude

and it comes with a new model, Opus 4.7…

ChatGPT's Nano Banana
Briefing

ChatGPT's Nano Banana

testing popular design tools…

Builders
Briefing

Builders

GPT-5.5 is a good model…

Elon doubled limits
Briefing

Elon doubled limits

Free ChatGPT got instantly better.…

Google's take on openclaw
Briefing

Google's take on openclaw

it's Anthropic's time for the mandate of heaven…

Opus 4.8
Briefing

Opus 4.8

NVIDIA and Microsoft birthed a new computer…

Anthropic study shows AI needs hours, not weeks, to build exploits from security patches
Briefing

Anthropic study shows AI needs hours, not weeks, to build exploits from security patches

Anthropic's security team found that its Mythos Preview AI model can turn security patches for Firefox and the Windows kernel into working exploits within hours…

OpenAI's IPO slips as Altman tells staff to expect a public offering "within the next year"
Briefing

OpenAI's IPO slips as Altman tells staff to expect a public offering "within the next year"

Sam Altman told employees he expects an OpenAI IPO "within the next year," but a delay to 2027 is possible. He frames it as caution around self-improving AI, th…

Google's new open model DiffusionGemma generates text from noise instead of word by word
Briefing

Google's new open model DiffusionGemma generates text from noise instead of word by word

Google released DiffusionGemma, a 26-billion-parameter model that generates text not token by token but through diffusion, similar to how image AI turns noise i…

Claude Fable 5: Anthropic admits "wrong tradeoff" after invisibly throttling rival AI researchers
Briefing

Claude Fable 5: Anthropic admits "wrong tradeoff" after invisibly throttling rival AI researchers

Anthropic reverses course on a policy that would have secretly undermined AI researchers, but another point of contention persists. The article Claude Fable 5: …

New Claude model - Fable
Briefing

New Claude model - Fable

everyone's building with it.…

Dario Amodei's new essay reads like a Cold War playbook for the AI age
Briefing

Dario Amodei's new essay reads like a Cold War playbook for the AI age

Anthropic publishes a sweeping essay and two policy frameworks. The company calls for binding audits of frontier models and paints a picture of AI as a strategi…

OpenAI vs. Anthropic: A price war over API tokens is brewing
Briefing

OpenAI vs. Anthropic: A price war over API tokens is brewing

OpenAI is weighing token price cuts to win customers from Anthropic, the Wall Street Journal reports. The article OpenAI vs. Anthropic: A price war over API tok…

OpenAI buys Ona to push Codex toward long-running, autonomous coding tasks
Briefing

OpenAI buys Ona to push Codex toward long-running, autonomous coding tasks

OpenAI is acquiring Ona, previously known as Gitpod, a startup founded in Kiel, Germany in 2020 that specializes in AI agents and secure cloud development envir…

GPT-5: It Just Does Stuff
Practice

GPT-5: It Just Does Stuff

Putting the AI in Charge…

Mass Intelligence
Practice

Mass Intelligence

From GPT-5 to nano banana: everyone is getting access to powerful AI…

Three Years from GPT-3 to Gemini 3
Practice

Three Years from GPT-3 to Gemini 3

From chatbots to agents…

Claude Code and What Comes Next
Practice

Claude Code and What Comes Next

With the right tools, AI can accomplish impressive things…

Claude Dispatch and the Power of Interfaces
Practice

Claude Dispatch and the Power of Interfaces

We often lack the tools for the job, even if the AI is capable enough…

Sign of the future: GPT-5.5
Practice

Sign of the future: GPT-5.5

One impressive step on the curve…

Introducing Google AI Threat Defense to help you outpace the adversary
Practice

Introducing Google AI Threat Defense to help you outpace the adversary

aside_block <ListValue: [StructValue([('title', 'Summary of today’s news'), ('body', <wagtail.rich_text.RichText object at 0…

Nano Banana 2 and Nano Banana Pro are generally available, and already powering creative workflows
Practice

Nano Banana 2 and Nano Banana Pro are generally available, and already powering creative workflows

Organizations are unlocking entirely new ways to use image generation and editing across their industries. To drive next-generation experiences, businesses are …

Evolving Dataflow to process massive datasets for machine learning
Practice

Evolving Dataflow to process massive datasets for machine learning

Google created MapReduce more than 20 years ago to solve the scaling problems in data processing that the then young company was running into. The AI era that w…

Developer's guide to Gemini Enterprise and A2UI integration
Practice

Developer's guide to Gemini Enterprise and A2UI integration

If you've built a chatbot, you know this conversation: User: "Book a table for two tomorrow at 7pm." Agent: "Okay, for what day?" User: "Tomorrow." Agent: "What…