Pricing changes log
What changed, when, and who it affects.
Toolvel’s tracker re-reads the official pricing pages it monitors once a day. When one of those pages changes, an editor checks the change against the vendor’s own pages before it is listed here. This is a selected log of verified pricing and billing changes since late September 2026, not a complete history. Page redesigns and reworded feature lists are not counted.
Price Watch · weekly email
Get the week’s verified pricing changes from this log by email. One message a week; unsubscribe from any issue.
Verified changes
Newest first. Entries are sorted by the date the change took effect where the vendor states one; otherwise by the announcement or first-detected date shown in the Dates column. A current pricing page cannot prove what a page said earlier, so a “Before” line appears only where the vendor’s own announcement or Toolvel’s dated check of the official page recorded it.
| Tool | What changed | Who it affects and what to do | Dates | Source |
|---|---|---|---|---|
| Claude (Claude Code) | New: Claude Fable 5.2 and Mythos 5.2 launched November 9, 2026. Before (Claude API pricing, 5.1 models): $10 input and $50 output per million tokens. After: $8 input and $40 output; cache reads $0.20; 5-minute cache writes $10; 1-hour cache writes $16; batch $4/$20; fast mode $16/$80. Mythos 5.2 remains limited to verified organizations. After the launch offers, Fable on Pro, standard Team and standard seat-based Enterprise uses usage credits; Max, Ultra, premium Team and premium seat-based Enterprise include Fable for up to 50% of weekly limits. | API users of Fable or Mythos 5.2, and paid Claude or Claude Code users choosing Fable. Confirm the model ID (claude-fable-5-2) and Claude Code version 2.1.321 or later. The launch offer that includes Fable for up to 50% of weekly limits on Pro and standard Team seats ends November 16, 2026 at 11:59 pm PT. Official pages do not publish absolute weekly or five-hour Fable token amounts. Toolvel breakdown. | Announced and effective Nov 9, 2026. Verified Nov 11. | Announcement · API pricing · What’s new · Launch offers · Plan access |
| Windsurf (Devin Desktop) | Before (official Devin pricing page, Toolvel check on Nov 9, 2026): Teams listed as “Up to 200 users.” After: the Teams card and the comparison table list “Up to 100 users.” Seat prices did not change ($80/month for the team plan + $40/month per full dev seat). Devin’s self-serve billing docs still said up to 200 users on November 10; those docs point to the pricing page for authoritative pricing. | Teams customers and anyone choosing Teams versus Enterprise based on headcount. Official pages do not say whether the 100-user cap applies to existing teams, or name an effective date. If your team is near or above 100 users, confirm with Cognition. | First detected Nov 10, 2026. Announcement and effective dates not found. Verified Nov 10. | Devin pricing · Self-serve billing docs |
| Windsurf (Devin Desktop) | New: SWE-2 Flash (Medium, High, Max) in Devin Desktop and Devin CLI, not in Devin cloud. Free until December 4, 2026 at 18:00 UTC; afterward $0.10 input, $0.01 cache read, $0.125 cache write and $0.50 output per million tokens. Medium is available on Free; High and Max need a paid plan. Also: on both token-based plans, GPT-5.6 Sol standard-context rates rise on November 21, 2026 at 07:00 UTC from $4 to $5 input and $20 to $30 output per million tokens; Fast mode costs twice those rates. Seat plan prices did not change. | Desktop and CLI users choosing SWE-2 Flash, and anyone on token-based plans using GPT-5.6 Sol. Check the in-app selector for account-specific cost; plan for Sol rates after November 21 and Flash list rates after December 4. | First detected Nov 9, 2026. Announcement date not found. Flash free through Dec 4, 2026 18:00 UTC; Sol change effective Nov 21, 2026 07:00 UTC. Verified Nov 9. | Devin model docs |
| GitHub Copilot | New: published cloud sandbox rates — compute $0.000024 per second, memory $0.000003 per GiB-second, stopped-session snapshot storage $0.005 per GiB-month. Compute and memory draw from included compute usage first; local sandboxes are free; spending uses product- or SKU-level budgets, not Bundled AI credits. Billing start date is not stated in the billing docs (the October 28 changelog said November 2026). Copilot plan prices did not change. | Paid Copilot users who run cloud sessions, and admins setting budgets. Check included compute usage and product/SKU budgets before heavy cloud use; local sandboxing does not add cost. | Rates found and verified Nov 9, 2026. Billing start date not announced in the billing docs. | Sandbox billing docs · Changelog |
| Claude (Claude Code) | Before (Claude pricing page, Toolvel daily checks through Oct 29): Max from $100 a month with 5x or 20x Pro usage; the 20x tier cost $200. After: Max is a single $100 plan with 5x Pro usage every five hours and every week. Ultra, at $200 a month, replaces Max 20x with 10x Pro usage every week, no five-hour limit for verified accounts, and $200 a month in API credits. | Max 20x subscribers move to Ultra automatically with no price change, according to Anthropic’s help center. A temporary five-hour limit applies until Anthropic verifies the account; Settings > Usage shows no Current session bar once it is verified. Max subscribers who mainly hit the five-hour window can compare Ultra. | First detected Oct 30, 2026. Announcement and effective dates not found. Verified Oct 30. | Claude pricing · Ultra plan help |
| OpenAI Codex | Before (OpenAI Pro tier help and its September 29 email): Pro 200 subscribers who did not qualify for the carryover got a lower Work and Codex allowance, and eligible existing subscribers were to drop to it after October 29. After: the reduction is canceled; lower-allowance Pro 200 subscribers were doubled to the higher allowance, and Pro 100 and Pro 500 now include twice their previous Work and Codex allowance. Business Premium usage also increased. Prices did not change. | All ChatGPT Pro subscribers and Business Premium seats; the change applies automatically, with no need to resubscribe. Weekly GPT-6 Pro message limits in Chat did not double on Pro 100 or Pro 500. OpenAI published no new multiples, so measure a week of use before changing tiers. | Announced and effective Oct 29, 2026. Found and verified Nov 5 (not caught by the pricing-page tracker). | OpenAI release notes · OpenAI help: Pro tiers |
| GitHub Copilot | New: Copilot access, an enterprise billing model with no Copilot seat charge for users with a GitHub Enterprise Cloud license, or $10 per user a month as Copilot Standalone. It has the same features and models as Copilot Enterprise but no included AI credits, so all usage is billed from the first request. Business ($19) and Enterprise ($39) seat prices did not change. | Enterprises working with a GitHub or Microsoft account team; self-serve customers cannot choose it yet. Moving is voluntary, and one enterprise cannot mix it with Business or Enterprise seats. Before asking to switch, price your full monthly usage without an allowance, not only today’s overage. | Announced and available Oct 28, 2026 for account-team customers. Transitions take effect on the first day of a month. Found and verified Nov 5. | GitHub changelog · Eligibility and transition |
| GitHub Copilot | New: one Copilot cloud experience on github.com, GitHub Mobile and other clients for paid Copilot plans, with sessions running in cloud sandboxes. Before (GitHub announcement): the included Actions entitlement was a fixed number of standard runner minutes. After: one shared, dollar-based compute entitlement of the same value that covers GitHub-hosted runners, cloud sandboxes and, on Team and Enterprise Cloud, larger runners. Copilot plan prices did not change. | Paid Copilot users who run cloud sessions, and admins. Access follows the “Copilot in the cloud” policy. Sandbox billing begins in November 2026 and draws from the included Actions entitlement; usage beyond it follows existing rates and budgets, so check Actions budgets before heavy cloud use. | Announced and rolling out from Oct 28, 2026. Found and verified Nov 5. | GitHub changelog |
| ElevenLabs | Before (ElevenLabs pricing page, Toolvel checks Oct 4 to Oct 18): Starter advertised at $1 for the first month, marked through October 18. October 19 observation: no first-month offer was shown. October 26 observation: the $1 first month was displayed again, labeled “Until Nov 1”. October 30 observation: the $1 first month (83% off) is still shown, but no end-date label is visible. The regular monthly price remains $6. The reappearance date is unconfirmed. | Eligible Starter subscribers: confirm the $1 first charge and $6 renewal at checkout. Creator’s first-month price ($11, then $22) is still shown. | Reappearance observed Oct 26, 2026; effective date unknown. End-date label not visible at the Oct 30 check. Earlier offer marked through Oct 18 and absent at the Oct 19 check. | ElevenLabs pricing |
| Runway | Before (Runway pricing page, Toolvel check Oct 18): Max listed unlimited Wan 3.0 and Wan 3.0 Prime in Tool Mode until October 18. After: the pricing page no longer advertises that unlimited Wan offer. | Max subscribers who used Wan 3.0 in Tool Mode. Check the model’s current availability and credit cost in the app before generating. | Promotion marked until Oct 18, 2026. No longer shown when Toolvel checked Oct 19. | Runway pricing |
| Cursor | Before: Pro $20, Pro+ $60, Ultra $200 a month; Teams $40 per user. After: Free, Plus $8, Premium $30, Super $100 and Ultra $200 a month; team seats at $30, $100 or $200. Third-party model requests are billed separately on Premium and above. | New subscribers. Existing Pro, Pro+ and Ultra subscribers keep their legacy plans until they switch. Before switching, compare your monthly third-party model use with the separately billed rates. | Announced and effective Oct 12, 2026. Verified Oct 14. | Announcement · Cursor pricing · Legacy plans |
| ElevenLabs | Now: $0.08 (v4) and $0.04 (v4 Turbo) per 1,000 characters on the API pricing page, with no discount label. Toolvel’s earlier coverage recorded a launch discount on these two models advertised through October 12. | API users of the two models. Re-estimate API budgets at the standard rates. | Discount advertised through Oct 12, 2026. Standard rates confirmed Oct 13; rechecked Oct 15. | ElevenLabs API pricing |
| Gemini Code Assist | Before: new Standard and Enterprise subscriptions on sale. After: no new subscriptions. Existing ones stay supported until the contract ends, and licenses can be added until January 31, 2027. | New customers can no longer buy. Existing subscribers should plan a move before their final renewal or contract end. | Effective Oct 9, 2026. Verified Oct 12. | Google subscriptions notice |
| GitHub Copilot | Before: code reviews requested by licensed members were always billed to the member’s own entitlement, and failed when it ran out. After: organization owners can choose to bill the organization instead. Member billing stays the default; organization billing requires AI Credits paid usage. | Organizations whose members use Copilot code review. If reviews fail on members’ quotas, consider organization billing with a budget. | Announced and available Oct 8, 2026 (opt-in). Verified Oct 15. | GitHub changelog |
| Claude (Claude Code) | New: monthly API credits of $100 on Max 5x and $200 on Max 20x; Team gets $20 per Standard seat and $100 per Premium seat, pooled and capped at $500. They do not cover interactive Claude Code or extra usage. | Max and Team subscribers after seven days on the plan. Use them for API, Console and Agent SDK work; unused credits expire each billing cycle. Max 20x is now Ultra (see the October 30 entry); the $200 credit carries over. | Announced Oct 7, 2026, rolling out that week. Verified Oct 15; rechecked Oct 17. | Announcement · Claude help |
| GitHub Copilot | Existing Business and Enterprise customers paying by credit card or PayPal are charged upfront for all assigned seats, starting with their next billing cycle on or after October 1. Seat prices did not change, and removing a seat mid-cycle does not give a prorated refund. | Organizations on Business or Enterprise that pay by card or PayPal. Review seat assignments before your billing date. | Announced Aug 28, 2026. Effective from Oct 1 at each customer’s next billing cycle. Verified Oct 15. | GitHub changelog |
| OpenAI Codex | Before (Codex pricing page, Toolvel check Sep 29): ChatGPT Pro from $100 a month with 5x or 20x Plus limits. After: three tiers at $100, $200 and $500 a month; Ultrafast only on Pro 500. | Pro subscribers and heavy Codex users. A lower Pro 200 allowance and an October 29 cutoff for existing subscribers were announced, then canceled on October 29 (see that entry). | Effective Sep 29, 2026. Verified Oct 1; rechecked Oct 17 and Nov 5. | OpenAI help: Pro tiers · Codex pricing |
| Windsurf (Devin Desktop) | Before (Devin pricing page, Toolvel check Sep 29): free SWE-2 use listed through October 10. After: listed through October 16 on the pricing page; Devin’s model docs gave October 15. On October 18 the pricing page listed free SWE-2 use with no end date. On October 30 the model docs said SWE-2 is free for self-serve plans in Devin Desktop and Devin CLI through November 30, 2026, with list pricing afterward; Devin cloud sessions are excluded. | Pro subscribers using SWE-2. Plan for list pricing after November 30 unless Devin extends the offer; the pricing page itself still shows no end date. | Detected Sep 29, 2026 (no effective date published). Verified Oct 1; rechecked Oct 17, 18, 19 and 30. | Devin pricing · Model docs |
Coming up
Dated offers, migrations and promotions that vendors have already published. Vendors can extend or cancel them, so check the source before relying on a date.
| Date | Tool | What happens | Source |
|---|---|---|---|
| Oct 31, 2026 (date passed) | Lovable | Lovable’s docs described the temporary chat pricing and daily chat allowance as applying through October 31. On November 5 the docs still showed that date and had not published the terms that follow. Check the Chat option and Credits used in your workspace. | Lovable docs: Chat mode |
| No end date shown | ElevenLabs Starter | The $1 first-month offer (then $6/month) was labeled “Until Nov 1” on October 26; on October 30 no end-date label was visible. Eligibility is account-specific, so confirm the first charge at checkout. | Official terms |
| November 2026 (no day given) | GitHub Copilot | Billing for cloud sandbox usage behind Copilot cloud sessions begins, per the October 28 changelog. Published rates: compute $0.000024/sec, memory $0.000003/GiB-sec, stopped-session storage $0.005/GiB-month; draws from included compute usage first; local sandboxes free. The billing docs still do not name a start day. | Sandbox billing docs · Changelog |
| Nov 16, 2026, 11:59 pm PT | Claude (Claude Code) | The launch offer that includes Fable models for up to 50% of weekly limits on Pro plans and standard Team seats ends, along with the one included Fable limit reset for those seats. After that, Fable on those plans uses usage credits if they are turned on. | Launch offers |
| Nov 21, 2026, 07:00 UTC | Windsurf (Devin Desktop) | On both token-based plans, GPT-5.6 Sol standard-context rates change from $4 to $5 input and $20 to $30 output per million tokens. Fast mode costs twice those rates. | Model docs |
| Nov 30, 2026 | Runway | 25% off Kling 4.0 generations on Standard, Pro and Max ends: 89 instead of 118 credits for 5 seconds at 1080p until then. | Runway pricing |
| Nov 30, 2026 | Runway | Legacy Unlimited subscribers are switched to Max automatically. Monthly and annual subscriber options. | Runway help |
| Nov 30, 2026 | Windsurf (Devin Desktop) | Free SWE-2 for self-serve plans in Devin Desktop and Devin CLI runs through this date, according to the model docs; list pricing applies afterward. The pricing page shows no end date. | Model docs |
| Dec 4, 2026, 18:00 UTC | Windsurf (Devin Desktop) | SWE-2 Flash free period ends. Afterward Medium, High and Max cost $0.10 input, $0.01 cache read, $0.125 cache write and $0.50 output per million tokens. Medium stays on Free; High and Max need a paid plan. Desktop and CLI only. | Model docs |
| Dec 1, 2026 | Windsurf (Devin Desktop) | Introductory self-serve rates for Gemini 4 Argon end. The additional 50% discount on those rates ran through October 27. Check account-specific selector pricing. | Official terms |
| Dec 1, 2026 | GitHub Copilot | Promotional token rates for Gemini 4 Argon end. | GitHub docs |
| Dec 31, 2026 | Runway Enterprise | Last date in the published signing window for eligible net-new Enterprise bonus credits, extended from October 31. The onboarding period that earns the bonus also ends by December 31, so a later signing leaves less time. Minimum spend and agreement conditions apply; this is not a blanket free-service period. | Official terms |
| March 8, 2027, 11:59 pm PT | Claude (Claude Code) | Unused Fable limit resets from the Fable 5.2 launch offer expire for eligible Max, Ultra, premium Team and legacy premium Enterprise seats. | Launch offers |
Get changes by email
The weekly Price Watch summarizes each week’s verified changes. See past issues and subscribe →
How changes are detected and verified: How Toolvel checks prices.