Your ChatGPT plan now pays for coding in Windsurf, Devin and more

Sign in with ChatGPT routes model costs to your Plus or Pro sub across 16 tools. Also new: GPT-6.1 Sol defaults, a 300 tok/s tier, and a Decisions API.

Nowline OCT 1 12:00 PM banner

Top AI stories from the last hour

Top AI stories from the last hour

Copy markdown

  • Your ChatGPT sub now foots the API bill

    Sign in with ChatGPT, unveiled at DevDay on Sep 29 and already live in Windsurf, lets ChatGPT Plus and Pro users charge GPT model usage to their subscription across 16 partner tools including Devin, Notion and Vercel — no separate API key or billing to wire up.

  • GPT-6.1 Sol becomes your tools' default

    Cline v4.1.22 and GitHub Copilot CLI v1.0.90-6 both shipped Sep 30, switching defaults to — or adding selection of — GPT-6.1 Sol, which OpenAI prices at roughly a fifth of GPT-6 Sol. Check your model config before the next run so you know what's answering, and at what cost.

  • A 300 tok/s tier lands in Codex

    GPT-6 Astra Ultrafast runs up to 8x faster in Codex and 6x in the API — about 300 tokens/sec — at 6x the per-token cost, and Ampcode already exposes it as 'Plaid speed.' Worth it for latency-bound agents where wall-clock time beats token price.

  • OpenAI ships a Decisions API

    A limited-preview Decisions API returns real-time answers to user-defined questions with a finite set of options instead of free text — OpenAI's move into the 'System-One' decision-model space, callable now. Handy for routing, scoring and guardrail checks without parsing prose.

  • Pro 500: 25x the Plus allowance

    A new Pro 500 plan grants 25x a ChatGPT Plus tier's usage and bundles Ultrafast access, aimed at builders who keep slamming into Codex and agent limits. If quota resets are your real bottleneck, this is the lever to pull.