Agent Sandbox vs Anthropic Code Execution

Anthropic Code Execution costs $0.05 per vCPU-hour. Agent Sandbox runs containers. Anthropic Code Execution runs containers.

Updated Oct 15, 2026

Agent Sandbox

Hosted API that runs Python and shell code for AI agents.

Isolation: Container
Free tier: $10 credit with a Google sign-in
Max session: Not listed

Visit site Agent Sandbox profile

Anthropic Code Execution

Sandboxed Python and bash containers that Claude runs through the Anthropic API.

Isolation: Container
Free tier: 1,550 free hours per month
Max session: Not listed

Visit site Anthropic Code Execution profile

What is different

3 fields differ. The rest are the same or not listed.

Field

Agent Sandbox

Anthropic Code Execution

Free tier: Agent Sandbox: $10 credit with a Google sign-in; Anthropic Code Execution: 1,550 free hours per month
Billing: Agent Sandbox: Per second; Anthropic Code Execution: Bundled machine time
SDKs: Agent Sandbox: Python; Anthropic Code Execution: Python, TypeScript, Go, Java, Ruby

Full comparison

18 fields, at each provider's list price

Agent Sandboxagentsandbox.co

Anthropic Code Executionplatform.claude.com/docs/en/about-claude/pricingplatform.claude.com

Pricing

vCPU, per hour: Agent Sandbox: Not listed; Anthropic Code Execution: $0.05
RAM, per GiB-hour: Agent Sandbox: Not listed; Anthropic Code Execution: Included
Storage: Agent Sandbox: $0.0005/MB for output files; Anthropic Code Execution: Not listed
Free tier: Agent Sandbox: $10 credit with a Google sign-in; Anthropic Code Execution: 1,550 free hours per month
Paid plan from: Agent Sandbox: Pay as you go; Anthropic Code Execution: Not listed
Billing: Agent Sandbox: Per second; Anthropic Code Execution: Bundled machine time

Limits

Max session: Agent Sandbox: Not listed; Anthropic Code Execution: Not listed
Idle stop: Agent Sandbox: After 5 minutes; Anthropic Code Execution: Not listed
Largest sandbox: Agent Sandbox: Not listed; Anthropic Code Execution: 1 vCPU · 5 GiB
Concurrent sandboxes: Agent Sandbox: Not listed; Anthropic Code Execution: Not listed

Platform

Isolation: Agent Sandbox: Container; Anthropic Code Execution: Container
GPU: Agent Sandbox: Not listed; Anthropic Code Execution: Not listed
Snapshots: Agent Sandbox: Not listed; Anthropic Code Execution: Not listed
Pause and resume: Agent Sandbox: Not listed; Anthropic Code Execution:

Developer

SDKs: Agent Sandbox: Python; Anthropic Code Execution: Python, TypeScript, Go, Java, Ruby
Integrations: Agent Sandbox: Claude Code; Anthropic Code Execution: Not listed

Trust

License: Agent Sandbox: Not listed; Anthropic Code Execution: Not listed
BYOC: Agent Sandbox: Not listed; Anthropic Code Execution: Not listed

Hourly cost

One sandbox running for one hour

Size

Agent Sandbox

Anthropic Code Execution

1 vCPU · 2 GiB: Agent Sandbox: Not listed; Anthropic Code Execution: $0.050
2 vCPU · 4 GiB: Agent Sandbox: Not listed; Anthropic Code Execution: $0.100
4 vCPU · 8 GiB: Agent Sandbox: Not listed; Anthropic Code Execution: $0.200

Compute only, at list rates.

Open the calculator

Which one fits

Based on the fields above

Agent Sandbox fits when

You use Claude Code.

Anthropic Code Execution fits when

You need pause and resume.

More comparisons

See all comparisons

Agent Sandbox

Anthropic Code Execution

What is different

Full comparison

Pricing

Limits

Platform

Developer

Trust

Hourly cost

Which one fits

Agent Sandbox fits when

Anthropic Code Execution fits when

More comparisons

Related

Compare

Pricing

Browse