Cognition said on 9 October 2026 that users can connect a personal ChatGPT Go, Plus or Pro plan to Devin, so that GPT model usage in Devin sessions draws from the ChatGPT plan instead of from Devin’s own quota.
Devin’s documentation calls this token sharing. It covers only the GPT portion of Devin’s multi-model harness, it requires a self-serve Devin Pro, Max or Teams plan, and it runs through OpenAI’s Sign in with ChatGPT programme, whose directory lists 12 apps with ChatGPT plan usage, Devin among them.
Business Pill · SWITCHING COSTS
A one-minute explainer of switching costs: what it costs a customer to move from one supplier to another. It teaches the general idea only and says nothing about any company in this story.
The key insight: As we read it, the documented change is a billing split rather than a merger of products: the user keeps two plans, and Devin routes only the GPT share of each session to the ChatGPT one, while every other model in its harness stays on Devin’s own meter.
Cognition’s post on X read: “You can use any personal ChatGPT plan with Devin.” It continued: “Just connect your Go, Plus, or Pro plan to Devin and all GPT model usage will draw down from you plan’s quota” (sic).
The Devin documentation page the post links to puts it this way: “Link your ChatGPT Go, Plus, or Pro subscription to Devin so GPT model usage in your sessions is billed to your ChatGPT plan instead of your Devin quota.”
Setup runs through Devin’s Settings, Connections page: the user chooses Continue with ChatGPT, signs in to ChatGPT, grants Devin permission to use the plan and turns on a switch labelled Use your ChatGPT plan. The documentation says one switch covers Devin Cloud, Devin Desktop and Devin CLI.
The 12 apps OpenAI’s Sign in with ChatGPT directory lists under ‘Apps with ChatGPT plan usage’ as read on 10 October 2026, with Devin highlighted. Lovable is marked coming soon. The same page lists sign-in-only apps and open-source integrations separately.
Who Qualifies on Each Side
On the ChatGPT side, the Devin documentation lists Go, Plus and Pro subscriptions, and adds: “Free ChatGPT accounts don’t include usage that can be shared with Devin.”
On the Devin side, token sharing is available on the self-serve Pro, Max and Teams plans. The documentation says it is not available on the Free plan or on Enterprise accounts.
On Teams, plans are linked per user. The documentation states: “There is no shared team-wide ChatGPT plan, and one member’s linked plan is never used for another member’s sessions.”
OpenAI’s own pages word eligibility differently. Its Help Center article lists Go, Plus and Pro, as Cognition does. Its DevDay 2026 recap says “Plan usage is available to Plus and Pro users in participating tools,” and its partner directory page also names Plus and Pro.
Which plans can use token sharing, per Devin’s documentation as read on 10 October 2026: ChatGPT Go, Plus or Pro on one side and a self-serve Devin Pro, Max or Teams plan on the other. ChatGPT Free, Devin Free and Devin Enterprise accounts are not eligible.
Only the GPT Share of the Harness Moves
The documentation is explicit that Devin quota is still used: “Devin Cloud runs on a harness that combines many models, and only some of them are OpenAI models.” The GPT portion is billed to the ChatGPT plan, and the other models in the ensemble are still billed to the Devin quota.
The plan works with all GPT models available in Devin except fast and priority variants. In Devin’s Adaptive mode, which chooses models automatically, the documentation says the ChatGPT subscription is used only when Adaptive routes to a GPT model.
When the ChatGPT plan runs out, the documentation says sessions keep running: Devin bills the GPT portion to the Devin quota, then to on-demand credits once that quota is used up, and goes back to the ChatGPT plan when its usage resets.
How OpenAI Meters It
OpenAI’s Help Center says that when a user turns this on, eligible AI requests count toward the ChatGPT Work and Codex usage included in the plan. Users can set a weekly usage limit for each app in ChatGPT’s settings, as a percentage of their overall weekly usage.
The article adds: “An app limit is a cap, not a separate pool of usage.” OpenAI’s partner page says connecting an app does not add a new allowance, and that the app may have its own fees.
OpenAI also says that using a ChatGPT plan does not by itself give the app access to the user’s ChatGPT conversations or memories, and that no OpenAI API key is shared.
A Follow-Up to DevDay
Devin was named before Cognition’s post. OpenAI’s DevDay 2026 recap read: “Use your allowance across 16 partners including Cognition’s Devin, Notion, Vercel, T3, OpenClaw, and Dactyl, and control how much each can use.” A 5 October post on X by @thsottiaux also listed Devin among “partners using Sign in With ChatGPT”. The documents we read give no start date for Devin’s token sharing.
OpenAI’s partner directory, as read on 10 October, lists 12 apps with ChatGPT plan usage, one of them Lovable, marked coming soon. It separately lists 5 apps with sign-in only, including HubSpot and Canva, and 4 open-source integrations, including OpenCode and T3. By our count, 12 and 4 make 16, the number in the recap; we have not matched the two lists name by name.
OpenAI’s Help Center says the programme “launched with a limited set of partners”, and asks commercial partners to apply through an interest form.
The Structural Read
Two subscriptions sit underneath it. Devin’s documentation requires a ChatGPT Go, Plus or Pro plan and a self-serve Devin Pro, Max or Teams plan; Free tiers on either side and Devin Enterprise accounts are left out.
The meter moves model by model. The documentation says the GPT portion is billed only to the ChatGPT plan, other models in the ensemble stay on the Devin quota, and Adaptive mode draws on the ChatGPT plan only when it routes to a GPT model.
OpenAI keeps the controls on its side. Its Help Center says each app gets a weekly limit set as a percentage of the user’s overall weekly usage, and that the limit is a cap rather than a separate pool.
Devin documentation, ‘Use your ChatGPT plan in Devin’
“Devin Cloud runs on a harness that combines many models, and only some of them are OpenAI models.”
Three Implications
CHATGPT SUBSCRIBERS Under OpenAI’s Help Center rules, eligible requests from an app count toward the ChatGPT Work and Codex usage included in the plan, and the user can set a weekly cap for each app.
DEVIN TEAMS Plans link per user, the documentation says, so on a Teams account each member’s sessions draw only on that member’s own ChatGPT plan.
OTHER APP MAKERS OpenAI’s Help Center says it launched with a limited set of partners and points commercial partners to an interest form.
The framework puts it this way: “The surface, not the model, governs autonomy. The agentic surface — the harness — is becoming where value pools in the entire AI stack.”
As we read it, token sharing writes that split into a billing rule: Devin’s harness decides which model runs each step, and only the steps that land on a GPT model are charged to the user’s ChatGPT plan.
What Is Not Established
Cognition’s post and the Devin documentation give no price change, no count of users who have linked a plan, and no share of Devin usage that runs on GPT models. OpenAI’s Help Center says only that how quickly included usage is used depends on the app and the tasks run.
Tibo (@thsottiaux), whose X profile reads “Codex & ChatGPT @OpenAI”, quoted Cognition’s post on X with the line “Your ChatGPT subscription is now also a Devin subscription.” Devin’s documentation sets narrower terms: a Devin plan above Free is still required, and only GPT usage moves to the ChatGPT plan.
Business Engineer Framework
The Agentic Harness War
A Business Engineer framework on why the agentic surface, the harness, is where value pools in the AI stack.
Cognition’s 9 October post and Devin’s documentation, as read on 10 October, say a ChatGPT Go, Plus or Pro subscriber on a self-serve Devin Pro, Max or Teams plan can have the GPT portion of Devin sessions billed to the ChatGPT plan, with other models still billed to Devin. OpenAI’s DevDay 2026 recap had already named Devin among the partners for its plan-allowance programme.
94,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. On 10 October 2026 we read Devin’s documentation page on using a ChatGPT plan, OpenAI’s Help Center article and Sign in with ChatGPT partner page, and the Sign in with ChatGPT section of OpenAI’s DevDay 2026 recap, and we read the posts by Cognition and @thsottiaux on X through the X API. The partner counts are as the directory stood when we read it, and the sum of 16 is our arithmetic. Nothing here is a forecast, and nothing here is financial or investment advice.
Gennaro Cuofano is a tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. By day he is an executive at a tech startup. He is also a business analyst and former assistant controller, with an International MBA with an emphasis on corporate finance (2012). His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer (businessengineer.ai), his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. See how we source.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
