Agentic Index
Anthropic Claude Code vs Cognition (2026)
Anthropic Claude Code last verified ·Cognition last verified
Devin takes a ticket and works alone in a cloud sandbox until it opens a pull request, while Claude Code works beside a developer who steers it from the terminal, an IDE or Slack. That verdict is the Agentic Index coverage score, graded from each vendor's own published materials.
Both plan, write and test code across a repository and keep memory across sessions. Claude Code checks each action with a classifier in auto mode or waits for the developer's approval, and it runs only Anthropic's Claude models. Teams extend it with plugins that bundle skills, hooks and subagents, and can test each plugin in CI before it ships. Developers get it through a Claude subscription, from Pro at $17 a month billed annually up to Max at $100 and Ultra at $200. Cognition's Devin gets its own shell, browser and editor, tests its work through computer use and can send the reviewer a video of the change. DeepWiki indexes connected repositories into searchable wikis. Devin Desktop, formerly Windsurf, adds an editor running Cognition's SWE models. Devin Oncall triages incidents from PagerDuty, Slack or Teams, and Outposts run sessions on the customer's own machines. Devin is free to try, with Pro at $20 a month, Max at $200 and Teams at $80 a month plus $40 a seat. Choose Claude Code to keep an engineer steering every session. Choose Devin to delegate scoped tickets and incidents to an agent that reports back.
On the Agentic Index coding agent ranking, neither Anthropic Claude Code nor Cognition clears the bar, which asks for all five merge loop capabilities documented in full. Anthropic Claude Code does not document knowledge grounding and RAG in full; Cognition does not document testing, debugging and optimization in full. 3 of the 96 vendors in the lane clear it. See the coding agent ranking
This comparison is published by Agentic Index, an independent agentic AI vendor research platform. Anthropic Claude Code and Cognition are each graded against the same 14 capability Agentic Index taxonomy, from the vendor's own public materials under the Agentic Index verification standard, alongside 1,165 researched vendors. No vendor pays for placement and no vendor has reviewed this page. How this evidence is graded
Choose Anthropic Claude Code if
- An engineer should steer each session, approving the plan and checking risky actions as they happen.
- You want to package internal skills, hooks and subagents as plugins and gate them with evals in CI.
- Your company already pays for Claude, and the agent should come with those seats.
- Large migrations should fan out from one script to hundreds of subagents that report back.
- Agent data should stay under Anthropic's commercial terms, with zero data retention available for qualified Enterprise accounts.
Choose Cognition if
- Your backlog holds well scoped tickets in Jira, Linear or Slack that an agent can take to a pull request alone.
- Reviewers want a screen recording of each change working, from tests Devin ran itself with computer use.
- Production incidents from PagerDuty, Slack or Teams should be triaged to a root cause before an engineer is paged in.
- You want repositories indexed into living wikis that the agent and your engineers both search.
- Sessions must run inside your own VMs, containers or Kubernetes clusters through Outposts.
| Feature | A Anthropic Claude Code |
C Cognition |
|---|---|---|
| Action & orchestration | ||
|
Integrations & Tool Calling Ability to connect agents to real systems through native integrations, OAuth-authenticated actions, custom tools, APIs, webhooks, or MCP-compatible tools. |
||
|
Anthropic Claude CodeIntegrations & Tool Calling MCP servers connect the agent to tools, databases, APIs, issue trackers and monitoring systems at local, project and user scopes, with OAuth 2.0 sign in for remote servers, claude.ai connectors, and managed settings that allow or deny servers across an organization, alongside built in file, search, shell and web tools. Sourcecode.claude.com/docs/en/mcp and docs/en/chromeread 2026-10-24 |
||
|
CognitionIntegrations & Tool Calling Native integrations cover GitHub, GitLab, Bitbucket, Azure DevOps, Slack, Microsoft Teams, Jira, Linear, PagerDuty and Pylon, and MCP connects Devin to hundreds of external tools and data sources, including monitoring, databases and documentation systems. Sourcedocs.devin.ai/integrations/overviewread 2026-10-28 |
||
|
Workflow Orchestration Ability to sequence, branch, retry, route, and combine deterministic workflow nodes with autonomous agent steps. |
||
|
Anthropic Claude CodeWorkflow Orchestration Subagents are workers Claude spawns for delegated tasks, and agent teams let a lead agent supervise long running peer sessions that share a task list. Dynamic workflows go further. Claude writes a JavaScript script that runs dozens to hundreds of subagents in the background, up to 1,000 per run and 16 at a time by default, asks before a run starts (every run in Manual mode, the first launch in Auto mode), and the script can be saved and rerun as a command. Sourcecode.claude.com/docs/en/workflowsread 2026-10-24 |
||
|
CognitionWorkflow Orchestration Dynamic workflows are Python scripts Devin writes and runs to orchestrate a team of Devin agents, deciding which agents run in what order, fanning items through stages in parallel on separate VMs and passing structured results from one stage to the next. Devin also delegates large tasks to managed Devins, each in its own VM. Sourcedocs.devin.ai/work-with-devin/dynamic-workflows and advanced-capabilitiesread 2026-10-28 |
||
|
Triggers & Channel Coverage How agents wake up and where they work: schedules, webhooks, message events, CRM events, inbox events, chat, email, voice, and collaboration tools. |
||
|
Anthropic Claude CodeTriggers & Channel Coverage Cloud scheduled tasks, in research preview, run a saved prompt against chosen repositories on a schedule, on an HTTP POST to the task's own endpoint, or on GitHub pull request and release events, so they keep working with the laptop closed. GitHub Actions and GitLab CI/CD automate code review and issue triage in CI, GitHub Code Review reviews every pull request automatically, and the agent answers @Claude in Slack and runs from the terminal, desktop, IDEs, the web and headless scripts. Sourcecode.claude.com/docs/en/routines and docs/en/overviewread 2026-10-24 |
||
|
CognitionTriggers & Channel Coverage Automations fire on Slack, GitHub, GitLab, Linear, Jira, Pylon, PagerDuty, schedule and webhook events, so once a trigger is defined Devin handles each event as it arrives without anyone tagging it. Scheduled sessions add recurring work, and Devin Oncall picks up alerts, bugs and incidents to investigate, fix or mitigate. Sourcedocs.devin.ai/product-guides/automations and devin-oncallread 2026-10-28 |
||
| Knowledge & context | ||
|
Knowledge Grounding & RAG Ability to ground agent behavior in company data through document ingestion, retrieval, external knowledge APIs, semantic search, or RAG layers. |
||
|
Anthropic Claude CodeKnowledge Grounding & RAG The agent grounds itself in the customer's repository by reading and searching it directly on each run, with project instructions from CLAUDE.md or AGENTS.md and path scoped rules, and MCP connections reach the customer's databases, issue trackers and internal APIs. Context is gathered fresh on each run, and Anthropic does not describe an index, graph or embeddings layer that persists between runs. Sourcecode.claude.com/docs/en/memory and docs/en/mcpread 2026-10-24 |
||
|
CognitionKnowledge Grounding & RAG Connected repositories are indexed from the DeepWiki settings page into wikis with architecture diagrams, source links and summaries, kept current and searchable, and Ask Devin uses them with code search to find context in the customer's code. Sourcedocs.devin.ai/work-with-devin/deepwiki and onboard-devin/index-reporead 2026-10-28 |
||
|
Memory & State Persistence Ability to persist context across a run, conversation, workflow, user, team, or longer-term memory layer. |
||
|
Anthropic Claude CodeMemory & State Persistence Auto memory, on by default in local sessions, holds notes Claude writes itself as it works, such as corrections the user gives and approaches the user confirms. It is scoped per repository and shared across worktrees, loaded into every session through a MEMORY.md index, kept outside the transcript retention sweep until the user or Claude edits or deletes it, and browsable and editable through /memory. CLAUDE.md files are standing instructions the user writes, separate from this memory. Sourcecode.claude.com/docs/en/memoryread 2026-10-24 |
||
|
CognitionMemory & State Persistence Devin Memory saves each user's preferences, corrections, decisions and lessons about their repos as one line notes in a personal memory drive, a Git repository of Markdown files that every new session loads. Each note links back to the session that taught it. Dreaming, a background session about once a day, consolidates overlapping notes and removes stale ones no session has used. Users browse their memory under Customize and can tell Devin to forget a note or turn personal memory off, and admins set it on or off for an organization or the whole enterprise. Knowledge adds shared instructions that users approve, scoped from one repository up to the enterprise. Sourcedocs.devin.ai/product-guides/memory and knowledgeread 2026-10-28 |
||
| Control & trust | ||
|
Human Oversight & Guardrails Approval steps, consent checkpoints, escalation rules, structured guardrails, policy constraints, and pause/resume controls. |
||
|
Anthropic Claude CodeHuman Oversight & Guardrails Auto mode is the starting mode for interactive terminal and VS Code sessions from v2.1.283. A separate classifier model reviews each action before it runs and blocks what it judges unsafe, the user's ask and deny rules still apply, and an organization can turn auto mode off in managed settings. Manual mode starts read only and asks before edits, shell commands and network requests, with approve once or allow from then on and fail closed matching for unmatched commands. Plan mode researches and proposes changes without editing, and claude -p and Agent SDK runs start in Manual. Sourcecode.claude.com/docs/en/permission-modes and docs/en/securityread 2026-10-24 |
||
|
CognitionHuman Oversight & Guardrails Local Hands lets a cloud Devin session run commands and edit files on the user's own computer only with the user's approval for each tool call. After a pull request Devin offers to test the app and waits for a click unless pre approved testing is turned on. Knowledge Devin proposes is saved only after a user edits and approves it, AI guardrails screen messages and PR comments and can block them before Devin acts, security profiles apply reusable restrictions to network, MCP and git access, and Devin Review runs only as often as the customer sets. Sourcedocs.devin.ai local-hands, testing-and-recordings, knowledge, ai-guardrails and security-profilesread 2026-10-28 |
||
|
Security, Identity & Governance RBAC, SSO, auditability, encryption, least-privilege tool access, compliance posture, and data handling policy. |
||
|
Anthropic Claude CodeSecurity, Identity & Governance A SOC 2 Type 2 report and an ISO 27001 certificate sit in the Anthropic Trust Center. Claude Code adds permission modes, sandboxed shell commands with filesystem and network isolation, command injection detection, trust verification for new folders and MCP servers, keychain credential storage, managed settings for organization enforcement and a HackerOne disclosure program. Under commercial terms Anthropic does not train models on Claude Code code or prompts unless the customer opts in. Sourcecode.claude.com/docs/en/security and docs/en/data-usageread 2026-10-24 |
||
|
CognitionSecurity, Identity & Governance Independent auditors assess Cognition's security program at least annually under SOC 2 Type II and ISO 27001, and data is protected with TLS 1.2 or higher in transit and AES-256 at rest. Enterprise adds SAML and OIDC SSO, SCIM 2.0 provisioning, IdP group mapping, IP access lists and custom roles for fine grained RBAC. Sourcecognition.com/legal/security, docs.devin.ai custom-roles, scim and devin.ai/pricingread 2026-10-28 |
||
|
Observability & Auditability Traces, logs, execution histories, metrics, audit events, and debugging detail for production agent behavior. |
||
|
Anthropic Claude CodeObservability & Auditability OpenTelemetry exports metrics, events and optional traces, with a tool decision event for every permission accept or reject and a tool result event for every completed tool call, and OTEL_LOG_TOOL_DETAILS adds commands, MCP server and tool names and inputs. Telemetry endpoints set in managed settings override developer settings, and all operations in cloud sessions are logged for audit. Sourcecode.claude.com/docs/en/monitoring-usage and docs/en/securityread 2026-10-24 |
||
|
CognitionObservability & Auditability The session Progress tab unifies shell commands with outputs, code edits and browser activity, and users can jump to any point in the command history and open individual steps after the fact. Session Insights adds a timeline of key events and knowledge usage for completed sessions, guardrail events land in audit logs with a session link, and the Devin MCP server and API expose session events for search. Sourcedocs.devin.ai session tools, session insights, ai-guardrails and devin-mcpread 2026-10-28 |
||
|
Deployment & Data Residency Deployment modes and options, including SaaS, dedicated cloud, VPC, on-prem, hybrid, local runtime, and self-hosting. |
||
|
Anthropic Claude CodeDeployment & Data Residency Claude Code runs on the developer's own machine by default, and Remote Control keeps code execution and file access local. Cloud sessions run in isolated Anthropic managed VMs with network access limited by default, or an organization routes them, scheduled tasks included, to self hosted environments on its own infrastructure. Prompts and outputs travel over TLS 1.2 or later, and at rest the Anthropic API uses AES-256 disk encryption, while Amazon Bedrock, Google Cloud's Agent Platform and Microsoft Foundry apply their own keys, with customer managed keys available on the first two. Commercial data is kept 30 days by default, and Zero Data Retention is turned on per organization for qualified Claude for Enterprise accounts. Sourcecode.claude.com/docs/en/data-usage and docs/en/securityread 2026-10-24 |
||
|
CognitionDeployment & Data Residency Outposts, generally available on Pro, Max and Teams, run Devin sessions inside infrastructure the customer controls (its own VMs, containers, Kubernetes clusters or on premises machines), so command execution, file edits and repository access stay on customer machines while inference and planning stay in Devin's cloud. Enterprise also offers a Cognition hosted single tenant VPC reached over PrivateLink, and federal customers can run Devin Desktop and the CLI fully air gapped against their own model endpoints. Sourcedocs.devin.ai/cloud/outposts/overview, enterprise/deployment/overview and federal/air-gapped/overviewread 2026-10-28 |
||
| Solution readiness | ||
|
Prebuilt Agents, Templates & Packs Ready-made workflows, packaged employees, templates, blueprints, industry solutions, and role-specific agents that reduce time-to-value. |
||
|
Anthropic Claude CodePrebuilt Agents, Templates & Packs Plugins package skills, agents, hooks and MCP servers as one installable unit, and Anthropic's official marketplace is added the first time an interactive terminal session starts, browsable from the /plugin Discover tab or Claude Marketplace on the web, with a context cost estimate for each official plugin and named plugins such as commit-commands. Saved workflows and the bundled /deep-research workflow run as commands, and workflows can ship inside plugins. Sourcecode.claude.com/docs/en/plugins and docs/en/workflowsread 2026-10-24 |
||
|
CognitionPrebuilt Agents, Templates & Packs Cognition publishes the Devin official marketplace of plugins, mostly one per integration such as Linear, Notion, Datadog or Snowflake, each bundling skills and optionally rules, hooks, MCP servers and subagents, which administrators install or require for an organization. Playbooks add reusable prompt libraries, and Skills are SKILL.md procedures committed to a repository that every session follows. Sourcedocs.devin.ai/product-guides/plugins and skillsread 2026-10-28 |
||
| Platform extensibility | ||
|
Model Flexibility & Routing Ability to work across multiple foundation models, route tasks to different models, or let buyers bring their own providers and keys. |
||
|
Anthropic Claude CodeModel Flexibility & Routing Every model Claude Code runs is Anthropic's. The customer picks among Claude models per session or subagent and picks where they are served (the Claude API, Amazon Bedrock, Google Cloud's Agent Platform or Microsoft Foundry), but not the model maker, and Anthropic does not support routing Claude Code to non-Claude models through any gateway. Sourcecode.claude.com/docs/en/llm-gatewayread 2026-10-24 |
||
|
CognitionModel Flexibility & Routing Users pick the model from Anthropic, OpenAI, Google and Cognition models, including the SWE-2 coding model, plus open models such as DeepSeek, Kimi and GLM, with a --model flag, a /model command or a config default, and Enterprise teams restrict which models are available in Team Settings. Adaptive routes each task to a model automatically, and Fusion pairs a frontier lead model with a lower cost sidekick in Devin Desktop and the CLI. Sourcedocs.devin.ai/cli/models, adaptive and fusionread 2026-10-28 |
||
|
APIs, SDKs & MCP Extensibility Composability layer: stable APIs, SDKs, MCP tool consumption/serving, custom tools, and integration into internal systems. |
||
|
Anthropic Claude CodeAPIs, SDKs & MCP Extensibility The Claude Agent SDK gives the same tools, agent loop and context management that power Claude Code as a Python and TypeScript library, with hooks, subagents, MCP, permissions, sessions, skills and plugins, and the CLI can be driven from any language with -p and JSON output. Sourcecode.claude.com/docs/en/agent-sdk/overviewread 2026-10-24 |
||
|
CognitionAPIs, SDKs & MCP Extensibility The Devin API is documented across more than 300 reference pages, with an Outposts API and CLI reference and a CLI plugin format, giving programmatic control of sessions and the platform. The official Devin MCP server at mcp.devin.ai/mcp authenticates with a Devin API key and exposes tools that create, search and steer sessions, manage playbooks, knowledge and schedules, and list integrations. Sourcedocs.devin.ai llms.txt (API documentation) and work-with-devin/devin-mcpread 2026-10-28 |
||
|
Testing, Debugging & Optimization Testing, debugging, scoring, retries, fallbacks, quality gates, and optimization loops for improving agent workflows before and after deployment. |
||
|
Anthropic Claude CodeTesting, Debugging & Optimization claude plugin eval runs a suite of realistic prompts against a plugin and scores the results, grading each case with a regex over the reply, a check that a particular tool was called, or a rubric a second model judges. It compares the run against a no plugin baseline and gates CI on the score, so a team can catch regressions when its plugin changes or a new model ships. Sourcecode.claude.com/docs/en/plugin-evalsread 2026-10-24 |
||
|
CognitionTesting, Debugging & Optimization After a pull request, Devin can enter a testing mode that plans one end to end flow, runs it in the desktop and sends a screen recording as proof, and Devin Review checks pull requests. Session Insights reviews a completed session with feedback and an improved prompt. These check the agent's output or one session after the fact. No harness scores the agent's behavior across runs, and FrontierCode is Cognition's own published benchmark. Sourcedocs.devin.ai testing-and-recordings, session-insights and devin-reviewread 2026-10-28 |
||
| Specialist automation | ||
|
Browser & Computer Use Browser, desktop, or remote/local computer control for workflows that cannot be handled through stable APIs alone. |
||
|
Anthropic Claude CodeBrowser & Computer Use The Chrome integration lets the agent work in a visible browser window with the user's logins, navigating, clicking, typing, filling forms, uploading files, reading console errors and DOM state and saving screenshots, in Chrome, Edge and other Chromium browsers. Site permissions come from the extension, and the agent pauses for logins and CAPTCHAs. Computer use extends this to native macOS apps. Sourcecode.claude.com/docs/en/chromeread 2026-10-24 |
||
|
CognitionBrowser & Computer Use Each session includes an interactive browser Devin operates for testing and visual verification alongside its shell and IDE, and a full desktop for Computer Use with screenshots and recordings. In testing mode Devin runs the app, clicks through one end to end flow in the browser and records the screen. Sourcedocs.devin.ai devin-session-tools, computer-use and testing-and-recordingsread 2026-10-28 |
||
Pricing snapshot
Sourced from the Index pricing dataset · open each vendor's profile for full detail.
| Pricing | A Anthropic Claude Code |
C Cognition |
|---|---|---|
|
Entry price Lowest public entry point |
From $17 per month for Claude Pro billed annually, or $20 monthly. | Free tier, Pro at $20 a month, Max at $200 a month, Teams at $40 a month per full seat with an $80 monthly minimum, and Enterprise at custom pricing. |
|
Pricing confidence How public the numbers are |
Public, exact | Public, exact |
|
Billing Primary billing axis |
Subscription per user (Pro, Max) or per seat (Team, Enterprise), each with plan usage limits, or per token on the Claude API. | Monthly plans with an included usage quota, plus prepaid on demand credits for usage beyond it. Teams charges $40 a month per full seat with an $80 monthly minimum, and flex seats are free. |
|
Variable cost Workload / overage exposure |
High variable cost | High variable cost |
|
Free tier / trial Try before you buy |
No free tier
|
Free tierTrial
|
|
Buying motion Self-serve vs sales call |
Self-serve | Self-serve |
|
Pricing verified When this pricing was last checked |
Product updates
Newest first, each linked to its source.
Anthropic Claude Code
-
· Human approval / guardrails
A new permission setting in Claude Code 2.1.319 denies any tool call that would otherwise wait on a permission prompt and lets the turn carry on, in terminal sessions, subagents and teammates.
Source -
· Security / enterprise
MCP server blocklists in Claude Code 2.1.318 now reach further.
Source -
· Observability / auditability
Claude Code 2.1.317 adds a cost report to the Claude apps gateway admin API, returning past spend over a date range in total, per developer, per day or per IdP group.
Source
Cognition
-
· Agent capability
Repeat reviews of the same pull request in Devin Review 2.0 now cover only the parts that changed since the last pass, by default on every PR.
Source -
· Agent capability
SWE-2 Flash, a new coding model trained by Cognition on top of Inkling-Small from Thinking Machines, is available now in Devin Desktop and Devin CLI.
Source -
· Agent capability
Code scans in Devin can now cover up to 1,000 repositories in one run, up from 500, in both the web app and the API.
Source
Claude Code runs only Anthropic's models. Cognition prices Enterprise by quote and bills features such as Automations, Deep Mode and Devin Review by usage, so model the credits a busy team would use. Neither vendor states how many tasks a plan covers, so trial both on real tickets.
More comparisons with Anthropic Claude Code or Cognition
Other matchups in coding agents
Not the pairing you were after? These compare a different set of coding agents on the same 14 capabilities.