OpenAI and Anthropic add spending limit features to their SDKs
OpenAI released a new SpendControl type in the openai‑python SDK version v3.30.0 on October 14 2026. The type lets developers set a per‑session spending cap expressed in USD cents, with a minimum of 1 and a maximum of 4,503,599,627,370,495 cents. The default is unlimited, and removing the limit does not reset the amount already recorded.
Key points
- OpenAI added SpendControl to openai‑python v3.30.0 (Oct 14 2026) for per‑session USD‑cent limits
- Anthropic added read‑only BetaSpendLimitUsageTierScope to anthropic‑sdk‑python v1.15.0 (Oct 15 2026) exposing monthly caps
- Neither SDK guarantees a session will stop when the configured limit is reached
Anthropic introduced a read‑only BetaSpendLimitUsageTierScope in anthropic‑sdk‑python version v1.15.0 on October 15 2026. It reports the monthly spend cap that the organization’s usage tier or a manually raised cap provides, expressed as a decimal string of cents (e.g., "50000" for $500). Only Claude Enterprise organizations can write limits; Console users can only read them.
Both SDKs note that the limits are best‑effort or informational and do not guarantee that a session will stop when the cap is hit. Developers are advised to treat the values as guidelines and implement their own monitoring if they need hard stop behavior.
Limits to stop AI overspending were added to OpenAI and Claude SDKs in the same week
note.com · 17 October 2026
Loading the full article…
This text was published by note.com and written by アライ｜AI思考実験とデータ分析. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Agents & Tools
All →- OpenAI's GPT-6 Pro remains unconfirmed as AI parallel processing concept explored · 1 src
- Masakazu tanaka's AI 'CurrentCut' wins japan's only award at Google hackathon · 1 src
- Perplexity open-sources SPACE, its AI agent sandbox platform, with partners CoreWeave, Akamai, and Crusoe · 3 src
- Anthropic’s Claude Code team cuts human code reviews by 40% with AI checks · 2 src
- Claude Code team says AI review now precedes human checks · 1 src
Comments
via GitHub Discussions