Cua Fleets vs Superserve

Cua Fleets costs $0.044625 per vCPU-hour and Superserve costs $0.0504 per vCPU-hour. Cua Fleets runs gVisor. Superserve runs Firecracker microVMs and has no session limit.

Updated Oct 14, 2026

Cua Fleets

Hosted Linux and Windows sandboxes for computer-use agents.

Isolation: gVisor
Free tier: Not listed
Max session: Not listed

Visit site Cua Fleets profile

Superserve

Firecracker microVM sandboxes that pause and resume.

Isolation: Firecracker microVM
Free tier: Not listed
Max session: No limit

Visit site Superserve profile

What is different

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

Field

Cua Fleets

Superserve

vCPU, per hour: Cua Fleets: $0.044625; Superserve: $0.0504
RAM, per GiB-hour: Cua Fleets: $0.0223125; Superserve: $0.0162
Storage: Cua Fleets: Free; Superserve: $0.000108 per GiB-hour
Largest sandbox: Cua Fleets: 8 vCPU · 32 GiB; Superserve: 2 vCPU · 2 GiB
Isolation: Cua Fleets: gVisor; Superserve: Firecracker microVM
Pause and resume: Cua Fleets: No; Superserve:
SDKs: Cua Fleets: Python, TypeScript, Rust; Superserve: Python, TypeScript
Integrations: Cua Fleets: OpenAI Agents SDK, MCP; Superserve: Claude Agent SDK, LangChain, OpenAI Agents SDK, MCP

Full comparison

18 fields, at each provider's list price

Cua Fleetscua.ai/#pricing

Superservesuperserve.ai/pricingsuperserve.ai

Pricing

vCPU, per hour: Cua Fleets: $0.044625; Superserve: $0.0504
RAM, per GiB-hour: Cua Fleets: $0.0223125; Superserve: $0.0162
Storage: Cua Fleets: Free; Superserve: $0.000108 per GiB-hour
Free tier: Cua Fleets: Not listed; Superserve: Not listed
Paid plan from: Cua Fleets: Not listed; Superserve: Pay as you go
Billing: Cua Fleets: Per second; Superserve: Per second

Limits

Max session: Cua Fleets: Not listed; Superserve: No limit
Idle stop: Cua Fleets: Not listed; Superserve: Not listed
Largest sandbox: Cua Fleets: 8 vCPU · 32 GiB; Superserve: 2 vCPU · 2 GiB
Concurrent sandboxes: Cua Fleets: Not listed; Superserve: Not listed

Platform

Isolation: Cua Fleets: gVisor; Superserve: Firecracker microVM
GPU: Cua Fleets: Not listed; Superserve: Not listed
Snapshots: Cua Fleets: Not listed; Superserve:
Pause and resume: Cua Fleets: No; Superserve:

Developer

SDKs: Cua Fleets: Python, TypeScript, Rust; Superserve: Python, TypeScript
Integrations: Cua Fleets: OpenAI Agents SDK, MCP; Superserve: Claude Agent SDK, LangChain, OpenAI Agents SDK, MCP

Trust

License: Cua Fleets: Not listed; Superserve: Not listed
BYOC: Cua Fleets: Your AWS, Google Cloud, or Modal account; Superserve:

Hourly cost

One sandbox running for one hour

Size

Cua Fleets

Superserve

1 vCPU · 2 GiB: Cua Fleets: $0.089; Superserve: $0.083
2 vCPU · 4 GiB: Cua Fleets: $0.179; Superserve: $0.166
4 vCPU · 8 GiB: Cua Fleets: $0.357; Superserve: $0.331

Compute only, at list rates.

Open the calculator

Which one fits

Based on the fields above

Cua Fleets fits when

You need a sandbox larger than 2 vCPU · 2 GiB.
You need SOC 2.

Superserve fits when

You want each sandbox in its own microVM.
You use the Claude Agent SDK or LangChain.
You want the lower list price: $0.166 against $0.179 an hour for 2 vCPU and 4 GiB.
You need snapshots.

More comparisons

See all comparisons

Cua Fleets

Superserve

What is different

Full comparison

Pricing

Limits

Platform

Developer

Trust

Hourly cost

Which one fits

Cua Fleets fits when

Superserve fits when

More comparisons

Related

Compare

Pricing

Browse