TL;DR
5 takeawaysOpenAI makes the Decisions API generally available: built with GPT-6 Luna for bounded answers, up to 10x faster than Luna via the Responses API, at $0.10 per million input tokens with no output-token charges
Sources: OpenAI
SpaceXSI introduces the SpaceX Super Intelligence Partner Network for companies building with Grok, saying more than 130 companies took part in early access
Sources: SpaceXSI
Anthropic donates Agent Skills to the Agentic AI Foundation under the Linux Foundation, saying more than 70 agent platforms have integrated the standard
Sources: Anthropic
Short: NVIDIA adds Nemotron 3.5 Embed, Rerank and Parse retrieval models
Sources: NVIDIAHugging Face
Short: Anthropic opens a research preview of Claude Duet
Sources: Anthropic
The Details
Story 01 / 03
OpenAI’s Decisions API reaches general availability
- What happened
- OpenAI made the Decisions API generally available on October 22. Built with GPT-6 Luna, it answers questions with bounded answers, returning a probability, a choice from supplied options, or a rubric score, from text, images or native JSON inputs. OpenAI says it is up to 10x faster than using GPT-6 Luna through the Responses API. Since the public beta, OpenAI added a native GPT-Live integration, prompt caching, US and EU inference residency, and up to 400 questions per request, twice the previous limit.
- Why it matters
- Pricing is $0.10 per million input tokens and $0.01 per million cached input tokens, with no output-token charges, according to OpenAI. Customers quoted by OpenAI include Box (about 5x faster route selection in its evaluation), Datadog (40x faster p95 than available frontier models), Ramp (4x faster and 30% cheaper than GPT-6 Luna) and Sierra (under 280 ms p95).
- Enterprise impact
- Routing, classification and risk checks that currently call a general model can be moved to a cheaper, faster bounded-answer endpoint.
- Startup impact
- Agent builders can put fast decision calls in the hot path, such as choosing the next action in a live conversation.
- What to focus on
- How developers compare it with small classifier models they run themselves.
Discuss this story
Comment with your GitHub account. One thread per story.
Story 02 / 03
SpaceXSI launches a partner network for Grok
- What happened
- SpaceXSI introduced the SpaceX Super Intelligence Partner Network on October 22, its official program for companies building products, solutions and services with Grok. It has two partner types: Transformation Partners, who help customers deploy the Grok platform across the enterprise, and Marketplace Partners, who build plugins and integrations so the Grok product suite works with existing tools.
- Why it matters
- SpaceXSI says more than 130 companies took part in the early-access phase and that engineers at more than 30 organizations have been trained through its Forward Deployed Engineer accreditation program. Partners quoted include Bain & Company, Databricks, NVIDIA, Snowflake, MongoDB, Sentry, Twilio and Zscaler.
- Enterprise impact
- Companies adopting Grok, Grok Bot, Grok Build or Cursor now have an official list of accredited delivery and integration partners.
- Startup impact
- Tool vendors can apply as Marketplace Partners to ship plugins alongside Grok.
- What to focus on
- Which plugins and integrations appear in the marketplace first.
Discuss this story
Comment with your GitHub account. One thread per story.
Story 03 / 03
Anthropic hands Agent Skills to the Agentic AI Foundation
- What happened
- Anthropic said on October 22 it is donating Agent Skills, the open standard in which a folder with a SKILL.md file teaches an agent a task, to the Agentic AI Foundation (AAIF) under the Linux Foundation. Anthropic co-founded the AAIF with Block and OpenAI and previously donated the Model Context Protocol to it. The donation includes the name and logo, the GitHub organization and the agentskills.io domain.
- Why it matters
- Anthropic says more than 70 agent platforms, including ChatGPT, Cursor, GitHub Copilot and VS Code, have integrated Agent Skills, and that 1 million skills have been published and nearly 280 million installed. Engineers from Anthropic, Google, OpenAI and other companies will maintain the standard as individuals.
- Enterprise impact
- Skills written once can be reused across agent platforms under vendor-neutral governance.
- Startup impact
- A neutral home lowers the risk of building products on top of the standard.
- What to focus on
- The proposed roadmap: versioned releases, a formal proposal process and security guidance for skill authors.
Discuss this story
Comment with your GitHub account. One thread per story.
Worth a Glance
NVIDIA Nemotron 3.5 Embed, Rerank and Parse
NVIDIA released open retrieval models, including a 631M-parameter document parser, and says adding Rerank cut agent token cost per query by about 14% with Embed 8B.
NVIDIA / Hugging Face(opens in a new tab)Claude Duet research preview
Anthropic’s proactive agent app sends a morning brief across connected apps and works on its own cloud computer; access is by waitlist, starting with US Pro and Max subscribers.
Anthropic(opens in a new tab)
One Thing Not to Miss Today
The agent stack is standardizing
Cheap decision endpoints, partner marketplaces and a neutral home for skills all point to agents becoming shared infrastructure rather than one-off builds.