# Devin vs Claude Code (November 2026): prices, models and workflow

> Both start at $20 a month, but Devin is a cloud agent that runs models from several labs, while Claude Code starts in your terminal and runs Claude only.

URL: https://claudegpt.si/news/devin-vs-claude-code  
Published: 2026-11-18T14:59:35.518Z  
Author: ClaudeGPT Si Newsroom  
Section: Reviews  
Topics: Devin, Cognition, Claude Code, Anthropic, Coding agents, Comparison  
Publisher: ClaudeGPT Si

## Key takeaways

- Devin and Claude Code both start at $20 a month and top out at $200 a month for individuals, as of November 18, 2026.
- Devin's Pro plan includes models from OpenAI, Anthropic, Google and SpaceXAI, while Claude Code runs Claude models only and defaults to Opus 5.5.
- Claude Code has a $100 Max plan between Pro and Ultra, and is not included in Claude's Free plan; Devin has a free plan with a light quota.

Devin vs Claude Code comes down to where the work happens. Devin is Cognition's cloud agent: you hand it a ticket or a Slack thread and it works in its own machine, then opens a pull request. Claude Code is Anthropic's coding agent that starts in your terminal or editor and can send work to the cloud when you want it to. Both cost $20 a month to start and $200 a month at the top individual tier, as of November 18, 2026 ([Devin pricing](https://devin.ai/pricing), [Claude pricing](https://claude.com/pricing)).

The bigger difference is the models. Devin's paid plans include models from OpenAI, Anthropic, Google and SpaceXAI, plus Cognition's own SWE-2. Claude Code runs Claude models only, and on paid plans it defaults to Opus 5.5.

This comparison is based on the products' docs, pricing pages and published results, read on November 18, 2026. It is not a hands-on test.

## Devin vs Claude Code at a glance

| | Devin | Claude Code |
| --- | --- | --- |
| Company | Cognition | Anthropic |
| Where it starts | Cloud agent (web app, Slack, Teams, API), plus Devin Desktop and Devin CLI | Terminal CLI, plus VS Code, JetBrains, desktop app, web and mobile |
| Free plan (as of Nov 18, 2026) | Yes, light quota, limited models | No, Claude Code is not in Claude's Free plan |
| Cheapest paid plan (as of Nov 18, 2026) | Pro, $20/month | Pro, $20/month, or $17/month billed annually |
| Middle tier (as of Nov 18, 2026) | None | Max, $100/month, 5x Pro usage |
| Top individual plan (as of Nov 18, 2026) | Max, $200/month | Ultra, $200/month, 10x Pro usage a week |
| Models | OpenAI, Claude, Gemini, SpaceXAI, open-source models, SWE-2 | Claude only: Opus 5.5 by default, Sonnet 5.5, Haiku 5.5, Fable |
| Usage resets | Pro: daily and weekly. Max: weekly, no daily cap | Every 5 hours and every week. Ultra: no 5-hour limit for verified accounts |
| Past the limit | Extra usage at API pricing; bought credits never expire | API credits on Max ($100/month) and Ultra ($200/month); Enterprise pays API rates |
| Cloud sessions | Included from Pro (Devin Cloud) | Included from Pro; no separate charge for the cloud VM |
| Hand-off between local and cloud | `/handoff` in Devin CLI, both directions | `claude --cloud` to start in the cloud, `--teleport` to pull a session back |
| Team plan (as of Nov 18, 2026) | $80/month team fee plus $40/month per full dev seat | $25/seat/month standard, $125 premium (monthly billing) |

## How Devin and Claude Code differ in daily use

Devin starts from the delegated task. The [Devin docs](https://docs.devin.ai) call it "an autonomous AI software engineer that can write, run and test code." Their rule of thumb: if a task would take you about three hours, Devin can most likely do it. You give it a prompt in the web app, a Linear or Jira ticket, or a Slack or Microsoft Teams thread. It works in its own workspace with a shell, an editor and a browser, and you can watch or take over in the embedded IDE.

Claude Code starts at your keyboard. Its [docs](https://code.claude.com/docs/en/overview) describe "an agentic coding tool that reads your codebase, edits files, runs commands, and integrates with your development tools." You run `claude` inside a project, and it works on your files with your tools. The same engine runs in VS Code, Cursor, JetBrains, a desktop app and the browser, and your `CLAUDE.md` instructions, settings and MCP servers (connections to outside tools) follow it across all of them.

Each has moved toward the other. Cognition [launched Devin Desktop on June 2, 2026](https://cognition.com/blog/introducing-devin-desktop) as "the next generation of Windsurf," an editor with a command center for local and cloud agents. Devin also has a [local CLI](https://docs.devin.ai/cli.md) that "runs directly in your terminal." Claude Code, for its part, now runs [cloud sessions](https://code.claude.com/docs/en/claude-code-on-the-web.md) on Anthropic-managed virtual machines that keep going after you close your laptop.

So the products overlap more than their origins suggest. The default posture still differs. Devin assumes you send work away and review the pull request. Claude Code assumes you sit with it and sometimes send work away.

## What Devin costs (as of November 18, 2026)

These prices come from [devin.ai/pricing](https://devin.ai/pricing) and Devin's [self-serve plans doc](https://docs.devin.ai/admin/billing/self-serve.md), as of November 18, 2026:

| Plan | Price | What you get |
| --- | --- | --- |
| Free | $0 | Light quota, limited models, unlimited Tab completions and inline edits, up to 2 Free-mode sessions at a time |
| Pro | $20/month | Frontier models from OpenAI, Claude, Gemini and SpaceXAI, open-source models, free use of SWE-2, Devin Cloud, use your ChatGPT plan |
| Max | $200/month | Everything in Pro, Devin Bots, a much larger weekly allowance with no daily cap |
| Teams | $80/month plus $40/month per full dev seat | Devin Bots, multiplayer sessions, SSO, three cost-control tiers, up to 100 users |
| Enterprise | Custom | Dedicated deployment, SCIM, role-based access control |

Three details matter more than the sticker price:

- **The Pro allowance resets daily and weekly.** The docs say the daily allowance is more than a seventh of the weekly one, so you can keep working through a weekend. Max drops the daily cap.
- **Extra usage is billed at API pricing.** When the allowance runs out, you buy on-demand credits. The docs say "purchased credits never expire."
- **Teams has no per-member allowance.** The self-serve doc says the team's $80 is charged as prepaid credits and all usage draws from that shared pool. The pricing page adds $40 a month for each full dev seat, which gets its own quota.

The pricing page lists Devin Bots on Max and Teams but does not describe them, and Cognition had not published a post about them as of November 18.

## What Claude Code costs (as of November 18, 2026)

Claude Code comes with Claude's paid plans. These prices come from [claude.com/pricing](https://claude.com/pricing), as of November 18, 2026:

| Plan | Price | Claude Code usage |
| --- | --- | --- |
| Free | $0 | Not included |
| Pro | $20/month, or $17/month billed annually ($200 up front) | Included; usage limits every 5 hours and every week |
| Max | $100/month | 5x Pro usage, plus $100 a month in API credits |
| Ultra | $200/month | 10x Pro usage a week, no 5-hour limit for verified accounts, $200 a month in API credits |
| Team | $25/seat/month standard, $125 premium ($20 and $100 billed annually) | Included; premium seats get 5x the usage of standard seats |
| Enterprise | $20/seat/month billed annually, 20-seat minimum, plus usage at API rates | Included |

Two details matter:

- **Usage is shared.** Claude Code draws from the same limits as the rest of your Claude use. The [cloud docs](https://code.claude.com/docs/en/claude-code-on-the-web.md) say cloud sessions count toward those limits too, with "no separate charge for the cloud VM."
- **You can skip the subscription.** Claude Code can also sign in with an Anthropic Console account or use an API key, and the CLI, VS Code and JetBrains can run through third-party providers such as Amazon Bedrock. Cloud sessions need a claude.ai account.

Below $200, Claude Code has a step Devin lacks: Max at $100. Above $20, Devin's only individual option is the $200 Max plan.

## Models: Devin picks from several labs, Claude Code runs Claude

Devin is model-agnostic. Its Pro plan includes models from four providers, and the pricing page suggests "mini models like Haiku, GPT 5.2 Mini, and open source models like Kimi K2.5" to stretch the quota. That means Devin can run Claude models inside its own agent.

Cognition also trains its own coding models. On November 4, it [released SWE-2 Flash](https://cognition.com/blog/swe-2-flash), built on Thinking Machines' Inkling-Small. Cognition says it runs at three times the speed of SWE-2 and a fifth of the price. In Cognition's own table, Claude Haiku 5.5 scores higher than SWE-2 Flash on seven of eight benchmarks, for example 39.2% against 20.0% on Terminal-Bench 4.0. SWE-2 Flash is free in Devin Desktop and Devin CLI through November 30.

Claude Code's [model settings](https://code.claude.com/docs/en/model-config.md) list Claude models only. On Pro, Max, Ultra, Team and Enterprise it defaults to Opus 5.5. You can switch to Sonnet 5.5 for daily coding, Haiku 5.5 for simple tasks, or a Fable model for the longest and hardest jobs; no plan uses Fable by default.

This is the clearest trade-off. Devin lets you mix labs under one agent. Claude Code gives up that choice and is built around one family of models.

## Working in the cloud and handing off

Both tools now move a session between your machine and the cloud.

In Devin, the CLI's [`/handoff`](https://docs.devin.ai/cli/handoff) command packages the conversation and your current git branch and sends them to a cloud Devin session with its own VM. Run in a cloud session, it does the reverse: it fetches the pull request branch and starts a local session on it.

In Claude Code, `claude --cloud "task"` starts a new cloud session from your repository's GitHub remote, and each call runs in parallel. `claude --teleport` pulls a cloud session into your terminal. From the CLI the move is one-way: you can't push an existing terminal session to the cloud, though the desktop app can. Claude Code can also watch a pull request and fix CI failures and review comments on its own, once the Claude GitHub App is installed.

Devin reaches more chat tools out of the box: Slack and Microsoft Teams. Claude Code covers Slack, plus GitHub Actions and GitLab CI/CD.

## Which one fits

Pick Devin if your team wants to hand off whole tickets and review pull requests, wants models from several labs under one plan, or already uses Windsurf, now Devin Desktop.

Pick Claude Code if you want the agent in your own terminal and editor, are happy with Claude models, or want a plan between $20 and $200, since Max at $100 has no Devin equivalent.

If you are comparing Claude Code with other editors and agents, our [Kiro vs Claude Code](https://claudegpt.si/news/kiro-vs-claude-code) comparison covers another pairing, and [Cursor vs Devin](https://claudegpt.si/news/cursor-vs-devin) puts Devin against Cursor's editor and cloud agents.

## Frequently asked questions

### Is Claude Code free?

No. Claude's Free plan does not include Claude Code. It comes with Pro at $20 a month, or $17 a month billed annually, and with Max, Ultra, Team and Enterprise, as of November 18, 2026. You can also use it with an Anthropic Console account or an API key.

### Devin vs Cursor: which is cheaper?

Devin's individual plans cost $20 or $200 a month as of November 18, 2026. Our Cursor vs Devin comparison covers Cursor's plans and how the two differ: https://claudegpt.si/news/cursor-vs-devin.

### Can Devin use Claude models?

Yes. Devin's Pro plan includes frontier models from OpenAI, Claude, Gemini and SpaceXAI, and the pricing page names Haiku as a way to stretch the quota. Claude Code, by contrast, runs only Claude models.

### Devin vs Claude Code vs Codex: which can use OpenAI models?

Devin can: its Pro plan includes OpenAI models and lists "Use your ChatGPT plan" as a feature. Claude Code cannot, since it runs Claude models only. This comparison does not cover Codex itself.

## Sources

- [Devin plans and pricing](https://devin.ai/pricing), Cognition
- [Self-serve plans](https://docs.devin.ai/admin/billing/self-serve.md), Cognition
- [Introducing Devin](https://docs.devin.ai), Cognition
- [Devin CLI](https://docs.devin.ai/cli.md), Cognition
- [Devin CLI /handoff](https://docs.devin.ai/cli/handoff), Cognition
- [Introducing Devin Desktop](https://cognition.com/blog/introducing-devin-desktop), Cognition
- [Introducing SWE-2 Flash](https://cognition.com/blog/swe-2-flash), Cognition
- [Claude plans and pricing](https://claude.com/pricing), Anthropic
- [Claude Code overview](https://code.claude.com/docs/en/overview), Anthropic
- [Use Claude Code in the cloud](https://code.claude.com/docs/en/claude-code-on-the-web.md), Anthropic
- [Model configuration](https://code.claude.com/docs/en/model-config.md), Anthropic
