SF Compute Autoresearch Sandboxes vs Tensorlake Sandboxes
SF Compute Autoresearch Sandboxes bills memory only, at $0.04572 per GiB-hour. Tensorlake Sandboxes costs $0.07 per vCPU-hour. SF Compute Autoresearch Sandboxes runs microVMs. Tensorlake Sandboxes runs Firecracker microVMs, offers GPUs and has no session limit.
Updated
SF Compute Autoresearch Sandboxes
- Isolation
- MicroVM
- Free tier
- None
- Max session
- Not listed
Tensorlake Sandboxes
- Isolation
- Firecracker microVM
- Free tier
- Free plan, $0 forever
- Max session
- No limit
What is different
9 fields differ. The rest are the same or not listed.
SF Compute Autoresearch Sandboxes
Tensorlake Sandboxes
- vCPU, per hour
- SF Compute Autoresearch Sandboxes: Included
- Tensorlake Sandboxes: $0.07
- RAM, per GiB-hour
- SF Compute Autoresearch Sandboxes: $0.04572
- Tensorlake Sandboxes: $0.015
- Storage
- SF Compute Autoresearch Sandboxes: Sandbox snapshots $0.10 per GiB-month of deduplicated bytes, $0.05 per GiB-month in cold storage
- Tensorlake Sandboxes: $0.0002 per GiB-hour
- Free tier
- SF Compute Autoresearch Sandboxes: None
- Tensorlake Sandboxes: Free plan, $0 forever
- Billing
- SF Compute Autoresearch Sandboxes: Memory only
- Tensorlake Sandboxes: Active CPU time only
- Largest sandbox
- SF Compute Autoresearch Sandboxes: 4 vCPU · 8 GiB
- Tensorlake Sandboxes: 16 vCPU · 64 GiB
- Concurrent sandboxes
- SF Compute Autoresearch Sandboxes: Up to 4,096
- Tensorlake Sandboxes: Up to 1,000
- Isolation
- SF Compute Autoresearch Sandboxes: MicroVM
- Tensorlake Sandboxes: Firecracker microVM
- Integrations
- SF Compute Autoresearch Sandboxes: MCP, Claude Code
- Tensorlake Sandboxes: OpenAI Agents SDK
Full comparison
18 fields, at each provider's list price
SF Compute Autoresearch Sandboxesautoresearch
Tensorlake Sandboxestensorlake
Pricing
- vCPU, per hour
- SF Compute Autoresearch Sandboxes: Included
- Tensorlake Sandboxes: $0.07
- RAM, per GiB-hour
- SF Compute Autoresearch Sandboxes: $0.04572
- Tensorlake Sandboxes: $0.015
- Storage
- SF Compute Autoresearch Sandboxes: Sandbox snapshots $0.10 per GiB-month of deduplicated bytes, $0.05 per GiB-month in cold storage
- Tensorlake Sandboxes: $0.0002 per GiB-hour
- Free tier
- SF Compute Autoresearch Sandboxes: None
- Tensorlake Sandboxes: Free plan, $0 forever
- Paid plan from
- SF Compute Autoresearch Sandboxes: Pay as you go
- Tensorlake Sandboxes: Pay as you go
- Billing
- SF Compute Autoresearch Sandboxes: Memory only
- Tensorlake Sandboxes: Active CPU time only
Limits
- Max session
- SF Compute Autoresearch Sandboxes: Not listed
- Tensorlake Sandboxes: No limit
- Idle stop
- SF Compute Autoresearch Sandboxes: Not listed
- Tensorlake Sandboxes: After 10 minutes
- Largest sandbox
- SF Compute Autoresearch Sandboxes: 4 vCPU · 8 GiB
- Tensorlake Sandboxes: 16 vCPU · 64 GiB
- Concurrent sandboxes
- SF Compute Autoresearch Sandboxes: Up to 4,096
- Tensorlake Sandboxes: Up to 1,000
Platform
- Isolation
- SF Compute Autoresearch Sandboxes: MicroVM
- Tensorlake Sandboxes: Firecracker microVM
- GPU
- SF Compute Autoresearch Sandboxes: Not listed
- Tensorlake Sandboxes:
- Snapshots
- SF Compute Autoresearch Sandboxes:
- Tensorlake Sandboxes:
- Pause and resume
- SF Compute Autoresearch Sandboxes:
- Tensorlake Sandboxes:
Developer
- SDKs
- SF Compute Autoresearch Sandboxes: Not listed
- Tensorlake Sandboxes: Python, TypeScript
- Integrations
- SF Compute Autoresearch Sandboxes: MCP, Claude Code
- Tensorlake Sandboxes: OpenAI Agents SDK
Trust
- License
- SF Compute Autoresearch Sandboxes: Not listed
- Tensorlake Sandboxes: Not listed
- BYOC
- SF Compute Autoresearch Sandboxes: Not listed
- Tensorlake Sandboxes: AWS, GCP, Azure, CoreWeave, Nebius or bare metal
Hourly cost
One sandbox running for one hour
SF Compute Autoresearch Sandboxes
Tensorlake Sandboxes
- 1 vCPU · 2 GiB
- SF Compute Autoresearch Sandboxes: $0.091
- Tensorlake Sandboxes: $0.100
- 2 vCPU · 4 GiB
- SF Compute Autoresearch Sandboxes: $0.183
- Tensorlake Sandboxes: $0.200
- 4 vCPU · 8 GiB
- SF Compute Autoresearch Sandboxes: $0.366
- Tensorlake Sandboxes: $0.400
Compute only, at list rates.
Open the calculatorWhich one fits
Based on the fields above
SF Compute Autoresearch Sandboxes fits when
- You use MCP or Claude Code.
- You want the lower list price: $0.183 against $0.200 an hour for 2 vCPU and 4 GiB.
- You want one rate for memory, with CPU included.
Tensorlake Sandboxes fits when
- You need GPUs.
- You use the OpenAI Agents SDK.
- You want to start without paying.
- You need a desktop in the sandbox.
More comparisons
Agent Sandbox vs SF Compute Autoresearch Sandboxes- Agent Sandbox vs Tensorlake Sandboxes
Agent37 Cloud vs SF Compute Autoresearch Sandboxes- Agent37 Cloud vs Tensorlake Sandboxes
AgentBay vs SF Compute Autoresearch Sandboxes- AgentBay vs Tensorlake Sandboxes
- Alibaba Cloud Agent Sandbox vs SF Compute Autoresearch Sandboxes
- Alibaba Cloud Agent Sandbox vs Tensorlake Sandboxes