Claude Agent SDK in our research
All researchStudy
AI Agent Pricing Index, October 2026
Pricing
Usage-based
14 of the 149 agents we checked bill by usage.
Price checked 2 September 2026.
What we checked on Claude Agent SDK
Writes code
The agent calls the same file tools Claude Code uses, so it authors and edits code from a prompt.
An agent is an application that completes a task by planning its own steps and calling tools that read files, run commands, or edit code.
Debugs and fixes
The vendor's own quickstart is an agent that locates crash bugs in a file and edits the fix in.
Use the Agent SDK to build an AI agent that reads your code, finds bugs, and fixes them, all without manual intervention.
Writes tests
The agent-loop walkthrough has the agent updating and re-running a project's tests alongside its edits.
A complex task ("refactor the auth module and update the tests") can chain dozens of tool calls across many turns, reading files, editing code, and…
Reviews code
The CLI surface is documented as a reviewer over a diff piped in from git or gh.
You can wrap a non-interactive call in a script to use Claude as a project-specific linter or reviewer.
Whole-codebase work
Repo-scale work is handled by fanning out subagents and by automatic compaction; no whole-repo index is claimed.
A research-assistant subagent can explore dozens of files without any of that content accumulating in the main conversation.
Ships it
Bash gives it git and script execution, so a deploy command is reachable, but the SDK documents no deployment surface of its own.
Run shell commands, scripts, git operations
Runs autonomously
With no turn or budget cap the loop runs to completion without a human in it.
Without limits, the loop runs until Claude finishes on its own, which is fine for well-scoped tasks but can run long on open-ended prompts ("improve…
Multi-agent
A parent agent can define specialized subagents, run independent tasks concurrently, and receive their reports.
Subagents are separate agent instances that your main agent can spawn to handle focused subtasks.
Agent permissions
Applications can combine modes, allow and deny rules, hooks, and a runtime approval callback.
Use permission modes and rules to define what's allowed automatically, and the canUseTool callback to handle everything else at runtime.
Claude
Claude is the only model family; you can pin a specific Claude model per session.
Set it explicitly (for example, model="claude-sonnet-5") to pin a specific model or to use a smaller model for faster, cheaper agents.
GPT
Every documented model path is a Claude endpoint: the Anthropic API or a cloud provider's Claude endpoint.
The SDK needs outbound HTTPS to api.anthropic.com, or to your provider's regional endpoint when running on Amazon Bedrock or Google Cloud's Agent…
Open models
The enumerated model aliases are all Claude models; no open-weight option appears anywhere in the option reference.
Accepts an alias such as 'fable', 'opus', 'sonnet', 'haiku', 'inherit', or a full model ID.
Model choice
Developers can override a subagent’s model with a Claude alias or a full model identifier.
On the page: AgentDefinition configuration table, model field
In your editor
The SDK ships as a library you embed; the editor and terminal surfaces belong to Claude Code, which the same table routes interactive work to.
The SDK is available as a library for Python and TypeScript only.
On the command line
claude -p is documented as the SDK's CLI surface, with JSON and stream-JSON output for scripted callers.
It's available as a CLI for scripts and CI/CD, or as Python and TypeScript packages for full programmatic control.
In your pipeline
Bare mode exists for CI reproducibility, and exit codes plus init-event fields let a pipeline fail on a bad run.
Bare mode is useful for CI and scripts where you need the same result on every machine.
In a browser
You host the SDK yourself; the hosted surface is Managed Agents, which the vendor names as a separate product.
Hosted REST API, a separate product from the Agent SDK. Anthropic runs the agent and the sandbox.
Self-hosted
The SDK is documented for production deployment on infrastructure operated by the application developer.
This page covers self-hosting on your own infrastructure.
Open source
The product is governed by Anthropic commercial terms rather than an open-source license stated in its documentation.
Use of the Claude Agent SDK is governed by Anthropic's Commercial Terms of Service.
Licence
Proprietary; Anthropic’s commercial terms expressly cover using the SDK in products and services for customers.
Use of the Claude Agent SDK is governed by Anthropic's Commercial Terms of Service, including when you use it to power products and services that you…
SDK languages
The documented SDK offers first-party libraries in Python and TypeScript, with no other languages named here.
The Agent SDK gives you the same tools, agent loop, and context management that power Claude Code, programmable in Python and TypeScript.
Tracing and evals
The SDK exports OpenTelemetry traces, metrics, and log events to compatible hosted or self-hosted backends.
The Agent SDK can export this data as OpenTelemetry traces, metrics, and log events to any backend that accepts the OpenTelemetry Protocol (OTLP).
How it meters
API use is metered by model token consumption, while session cost estimates are calculated from token counts.
Claude Code charges by API token consumption.
Price per vCPU-hour
$0.05 an active CPU-hour on Code execution: it bills CPU time while working, not wall-clock vCPU time, so it does not compare directly.
Code execution Run Python code in a sandboxed environment for advanced data analysis. 50 free hours of usage daily per organization. Additional hours…
claude.com/pricingChecked
Price per 1,000 requests
$10 per 1,000 requests in Web search, its cheapest mode.
Web search Give Claude access to the latest information from the web. Doesn’t include input and output tokens required to process requests. Cost $10…
claude.com/pricingChecked
Free tier
A developer can create an Anthropic account, obtain an API key, and start without contacting sales.
Get an API key from the Claude Console, then set it as an environment variable in the shell where you'll run your agent:
API access
The product is itself a library for your stack, in Python or TypeScript, with a CLI fallback for other languages.
A library that runs the agent loop in your own process, in Python or TypeScript.
MCP server
The agent is a full MCP client over stdio, HTTP/SSE and in-process servers; the docs describe no way to expose your agent as an MCP server another agent could query.
MCP servers can run as local processes, connect over HTTP, or execute directly within your SDK application.
Bring your own key
Applications authenticate using their own Anthropic API key supplied through the process environment.
The SDK reads the key from the environment of the process that runs your agent; it doesn't load .env files automatically.
A company can buy it
Organizations can use the Claude Console as an API-first, pay-as-you-go route for Agent SDK workloads.
Claude Console | You're API-first or want pay-as-you-go billing
Seat model
The API the Agent SDK runs on is sold per token, not per seat: the pricing page's API tab lists only per-MTok model rates and per-use feature rates, with no seat price, seat floor or seat ceiling anywhere in it, and the per-seat figures on the page belong to the Claude chat subscriptions instead.
Fable 5.1 ... Input $10 / MTok Output $50 / MTok ... Managed Agents ... Cost $0.08 per session-hour for active runtime ... Web search ... Cost $10 /…
claude.com/pricingChecked
Pooled budget
One organization funds every key, with per-workspace monthly spend caps and alerts; the SDK also accepts a per-query maxBudgetUsd.
Use workspaces to separate different projects, environments, or teams while maintaining centralized billing and administration.
Admin controls
Organizations can enforce policies controlling tools, commands, MCP servers, and network destinations ahead of local settings.
Claude Code enforces organization policy through managed settings that take precedence over local developer configuration.
Audit log
OpenTelemetry can emit tool decisions, tool results, safety events, MCP connections, and permission mode changes.
These events become a per-user audit trail you can forward to a Security Information and Event Management (SIEM) platform:
Single sign-on
SSO is configured at the Claude account level, although this documentation does not name a protocol.
SSO, SCIM provisioning, and seat assignment are configured at the Claude account level.
Opt out of training
Anthropic states that code and prompts on Team, Enterprise, API, and cloud-provider plans are not used for training.
On Team, Enterprise, Claude API, and cloud provider plans, Anthropic doesn't train models on your code or prompts.
Data residency
First-party API inference is global by default, with a US-only inference option for supported newer models.
Global routing (the default) uses standard pricing.
Getting out
Session transcripts are written automatically to local disk and can be mirrored to shared storage, while credit-expiry terms are not stated here.
The SDK writes it to disk automatically so you can return to it later.
Certifications
Anthropic directs users to its Trust Center for SOC 2 Type 2 and ISO 27001 compliance materials.
Learn more and access resources (SOC 2 Type 2 report, ISO 27001 certificate, etc.) at Anthropic Trust Center.
Claude Agent SDK price and terms
What needs a higher plan
- Claude Enterprise
- Single sign-on
For AI agents
- MCP server
- Partly
- API
- Yes
Pricing page, captured
Claude Agent SDK alternatives
All 9 alternatives- This spot is openBe the first agent people see on Claude Agent SDK's page.
- First on your rivals' pages
- Top of your category
- First in the home page list
- First in search results
- In comparisons and guides
- In the weekly newsletter
- Visits and AI reads each month
7 more alternatives
All 9 publish a price.
- OpenAI Agents SDK (Python)FreeOpen-source Python framework for building agents with tools, handoffs, guardrails and tracing.
- Microsoft Agent FrameworkFreeOpen-source .NET, Python and Go SDK for building AI agents and graph-based multi-agent workflows.
- LangGraphFree, then $39 a seat a monthOpen-source orchestration framework and runtime for stateful, long-running LLM agents, from LangChain.
- CrewAIFreeOpen-source multi-agent framework with a managed platform for deploying and governing agent workflows.
Questions about Claude Agent SDK
- How much does Claude Agent SDK cost?
- Claude Agent SDK costs $0.10 per 1m input tokens. From platform.claude.com/docs/en/about-claude/pricing, checked 29 October 2026.
- Does Claude Agent SDK have an API or an MCP server?
- It has an API. It has a limited MCP server.
- Is Claude Agent SDK open source, and can you host it yourself?
- It is not open source. You can host it yourself.
- What have we checked on Claude Agent SDK?
- 39 facts, under these headings: What it does, how much it does unattended, which models it runs on, where you use it, whose machine it runs on, building with it, what it costs to run, buying it for a team, what happens to your code.
- How do we check Claude Agent SDK?
- We hold 39 facts on Claude Agent SDK, quoted from 28 pages. Facts and price last checked 29 October 2026. Last change found 18 October 2026.
Sources and changes
Last changed
Sources28
- code.claude.com/docs/en/agent-sdk/overview
- code.claude.com/docs/en/agent-sdk/subagents
- claude.com/pricing
- code.claude.com/docs/en/agent-sdk/agent-loop
- code.claude.com/docs/en/agent-sdk/quickstart
- platform.claude.com/docs/en/about-claude/pricing
- code.claude.com/docs/en/admin-setup
- code.claude.com/docs/en/agent-sdk
- code.claude.com/docs/en/agent-sdk/hosting
- code.claude.com/docs/en/headless
- code.claude.com/docs/en/agent-sdk/observability
- code.claude.com/docs/en/authentication
- platform.claude.com/docs/en/manage-claude/compliance-activity-feed
- code.claude.com/docs/en/model-config
- code.claude.com/docs/en/agent-sdk/claude-code-features
- code.claude.com/docs/en/agent-sdk/cost-tracking
- code.claude.com/docs/en/agent-sdk/mcp
- code.claude.com/docs/en/agent-sdk/permissions
- code.claude.com/docs/en/agent-sdk/secure-deployment
- code.claude.com/docs/en/agent-sdk/session-storage
- code.claude.com/docs/en/agent-sdk/sessions
- code.claude.com/docs/en/costs
- code.claude.com/docs/en/data-usage
- platform.claude.com/docs/en/manage-claude/admin-api
- platform.claude.com/docs/en/manage-claude/api-and-data-retention
- platform.claude.com/docs/en/manage-claude/data-residency
- platform.claude.com/docs/en/manage-claude/workspaces
- code.claude.com/docs/en/security
Written about it elsewhere1
What changed on this page1
- Price
Is this yours?
AI assistants like ChatGPT and Claude opened The Agents Index pages 2,746 times in September. Claiming is free. You get this page's reader numbers and the verified owner mark. You confirm it with an email at code.claude.com.
Report an error if a fact is wrong.