AI Usage Limits: change log

Anthropic: Claude's Enterprise seat article now spells out that each Enterprise seat belongs to one…

Fri, 16 Oct 2026 12:00:00 +0000

Claude's Enterprise seat article now spells out that each Enterprise seat belongs to one named person: no sharing logins, no using someone else's account to get around your own spend or usage limits, and seat reassignment is for normal team changes, not rotating one seat among several active users. Applies to every Enterprise seat type.

Date: Date not stated in source (seen 2026-10-16). Source: support.claude.com/en/articles/13393991

xAI: Grok Voice 3.0 (`grok-voice-3.0`) is available in the Voice Agent API at $0.08/min…

Fri, 16 Oct 2026 12:00:00 +0000

Grok Voice 3.0 (grok-voice-3.0) is available in the Voice Agent API at $0.08/min ($4.80/hr) plus $0.004 per text input, the same price as Grok Voice 2.0. grok-voice-latest now points to it automatically; grok-voice-2.0 is deprecated.

Date: Date not stated in source (seen 2026-10-16). Source: x.ai/news/grok-voice-3, docs.x.ai release notes, docs.x.ai pricing

Anthropic: If you upgrade your Claude plan after hitting a session or weekly limit, the new plan's…

Fri, 16 Oct 2026 12:00:00 +0000

If you upgrade your Claude plan after hitting a session or weekly limit, the new plan's limits apply right away; you don't have to wait for the reset. Upgrades take effect immediately, with unused time on the old plan credited.

Date: Date not stated in source (seen 2026-10-16). Source: support.claude.com/en/articles/11647753, claude.com/pricing

Google: Gemini 4 Argon is rolling out in Antigravity (desktop app and CLI) to paid Google AI Pro…

Thu, 15 Oct 2026 12:00:00 +0000

Gemini 4 Argon is rolling out in Antigravity (desktop app and CLI) to paid Google AI Pro subscribers, with a separate weekly-only Argon quota on top of the existing Gemini 3.8 Flash quota. Google hasn't published the quota size.

Date: 2026-10-15. Source: @antigravity on X, antigravity.google/blog

Anthropic: Claude Opus 5.5 fast mode on the Claude API is out of research preview and the waitlist…

Thu, 15 Oct 2026 12:00:00 +0000

Claude Opus 5.5 fast mode on the Claude API is out of research preview and the waitlist is gone: up to 2.5x faster output at $8 input / $40 output per 1M tokens. Also in Microsoft Foundry; Google Cloud "coming soon".

Date: 2026-10-15. Source: docs.anthropic.com pricing, @ClaudeDevs on X

Cursor: Cursor's model price list now includes Muse Wonder (Meta): $4 input, $0.20 cached input,…

Fri, 16 Oct 2026 12:00:00 +0000

Cursor's model price list now includes Muse Wonder (Meta): $4 input, $0.20 cached input, $20 output per 1M tokens, up to 1M tokens in Max Mode with no long-context surcharge. Muse Spark 1.3 is still listed but now hidden by default.

Date: Date not stated in source (seen 2026-10-16). Source: cursor.com/docs/models-and-pricing

Cursor: Cursor docs and help now lead with The X Subscription (Free, Plus $8, Premium $30, Super…

Thu, 15 Oct 2026 12:00:00 +0000

Cursor docs and help now lead with The X Subscription (Free, Plus $8, Premium $30, Super $100, Ultra $200; weekly pool for Cursor models shared with Grok; third-party models from prepaid credit on Premium+). Legacy Cursor plans (Hobby, Pro, Pro+, Ultra, Teams) are documented separately and keep two monthly pools plus on-demand usage. Switching is optional. Official pages no longer publish a dollar size for the legacy Other Models pool.

Date: Date not stated in source (seen 2026-10-15). Source: cursor.com/docs/models-and-pricing, cursor.com/help/account-and-billing/pricing.md, cursor.com/help/models-and-usage/usage-limits, legacy Cursor plans

Site: Added a Gemini usage limits guide (site update, not a Google change).

Thu, 15 Oct 2026 12:00:00 +0000

Added a Gemini usage limits guide (site update, not a Google change).

Date: 2026-10-15. Source: support.google.com/gemini/answer/16275805

Meta: Meta released open-weight Muse Spark Open (816B total / ~42B active parameters, 1M-token…

Thu, 15 Oct 2026 12:00:00 +0000

Meta released open-weight Muse Spark Open (816B total / ~42B active parameters, 1M-token context) under the Muse Model License. Weights on Hugging Face and in the Amazon SageMaker AI, Google Cloud Model Garden and Microsoft Foundry catalogs; hosted by API providers such as OpenRouter and Groq (Groq in limited preview). The announcements don't mention any change to Muse app or Muse Code plans or limits.

Date: 2026-10-15. Source: research.meta.ai/blog/muse-spark-open, dev.meta.ai/resources/blog/build-with-muse-spark-open

Anthropic: Claude Enterprise (usage-based / seat price + usage at API rates): $20 per seat per month…

Thu, 15 Oct 2026 12:00:00 +0000

Claude Enterprise (usage-based / seat price + usage at API rates): $20 per seat per month billed annually, plus usage at API rates. New Enterprise plans require a minimum of 20 seats.

Date: Date not stated in source (seen 2026-10-15). Source: claude.com/pricing, claude.com/solutions/enterprise, support.claude.com/en/articles/13393991

OpenAI: If you reach a Codex usage limit during an active turn, Codex can continue working on…

Thu, 15 Oct 2026 12:00:00 +0000

If you reach a Codex usage limit during an active turn, Codex can continue working on that turn, subject to fair-use limits (credit/allowance plans). On token-based Enterprise plans, continuation depends on the workspace budget and user limits.

Date: Date not stated in source (seen 2026-10-15). Source: help.openai.com/en/articles/11369540

Anthropic: Claude Code wrap-up allowance: if you hit the 5-hour session limit mid-response, Claude…

Thu, 15 Oct 2026 12:00:00 +0000

Claude Code wrap-up allowance: if you hit the 5-hour session limit mid-response, Claude Code may keep working briefly to finish the step in progress ("Usage limit reached · wrapping up"). Small and capped, counts toward the weekly limit, only for a response already in progress. Pro: up to once per weekly usage period. Max: each time a 5-hour limit is reached, within the weekly limit. Rolling out gradually.

Date: Date not stated in source (seen 2026-10-15). Source: support.claude.com/en/articles/11145838

OpenAI: GPT-5.5 retires from Codex with ChatGPT sign-in at noon PDT (19:00 UTC). API-key Codex…

Wed, 14 Oct 2026 12:00:00 +0000

GPT-5.5 retires from Codex with ChatGPT sign-in at noon PDT (19:00 UTC). API-key Codex and the OpenAI API are not affected. GPT-5.5 credit rates were removed from the Codex pricing page.

Date: 2026-10-14. Source: learn.chatgpt.com/docs/models, chatgpt.com/codex/pricing

X / Cursor: Pricing pages now say paid plans include Grok API access, billed separately at standard…

Wed, 14 Oct 2026 12:00:00 +0000

Pricing pages now say paid plans include Grok API access, billed separately at standard model rates and not drawn from the shared usage pool. The pool description no longer lists API access (our Oct 13 saved copy of cursor.com/pricing listed it).

Date: Date not stated in source (seen 2026-10-14). Source: x.ai/pricing, cursor.com/pricing

X / Cursor: The X Subscription launches (Plus $8, Premium $30, Super $100, Ultra $200) with one…

Tue, 13 Oct 2026 12:00:00 +0000

The X Subscription launches (Plus $8, Premium $30, Super $100, Ultra $200) with one weekly pool shared by Grok and Cursor. Existing plans continue.

Date: Date not stated in source (live as of 2026-10-13). Source: https://docs.x.ai/x-subscription/pricing

OpenAI: Global Codex usage reset for Plus, Pro and Business. Applied immediately, not banked.

Mon, 07 Sep 2026 12:00:00 +0000

Global Codex usage reset for Plus, Pro and Business. Applied immediately, not banked.

Date: 2026-09-07. Source: https://help.openai.com/en/articles/20001498-how-banked-codex-resets-work

OpenAI: Banked Codex resets granted to eligible Plus, Pro and Business accounts with the GPT-6…

Thu, 03 Sep 2026 12:00:00 +0000

Banked Codex resets granted to eligible Plus, Pro and Business accounts with the GPT-6 Astra rollout.

Date: 2026-09-03 / 04. Source: https://help.openai.com/en/articles/20001498-how-banked-codex-resets-work