Pricing
Pay for what your agent does. Nothing while it rests.
No seats, no minimums — prepaid credits on two meters, itemized to the turn. Or bring your Claude or ChatGPT subscription and model usage is free.
Harness
The agent software you choose at launch — swap anytime without losing its history.
Bring your subscription
Already pay Anthropic or OpenAI? Connect your plan once and eligible agents run on it — AgentSky bills $0 for their model usage.
Model usage
No subscription? Pay per turn for the tokens your agent actually consumes.
Published model rates
Model tokens are billed at the rates shown below. Compute, connectors, and paid capabilities use the separate prices published on this page, so a session total can include more than tokens.*
Claude Fable 5.1
Claude Opus 5
Claude Sonnet 5
Claude Haiku 4.5
GPT-5.6 Sol
GPT-5.6 Terra
GPT-5.6 Luna
GPT-6 Astra
DeepSeek V4 Pro
DeepSeek V4 Flash
Gemini 3.8 Flash
GLM-5.3
GLM-5.3-Flash
Kimi K3
Kimi K2.7 Code
Grok 4.6
| Model | Input / 1M tokens | Output / 1M tokens |
|---|---|---|
| Claude Fable 5.1 | $10.00 | $50.00 |
| Claude Opus 5 | $5.00 | $25.00 |
| Claude Sonnet 5 | $2.00 | $10.00 |
| Claude Haiku 4.5 | $1.00 | $5.00 |
| GPT-5.6 Sol | $5.00 | $30.00 |
| GPT-5.6 Terra | $2.00 | $12.00 |
| GPT-5.6 Luna | $0.20 | $1.20 |
| GPT-6 Astra | $10.00 | $50.00 |
| DeepSeek V4 Pro | $1.32 | $3.96 |
| DeepSeek V4 Flash | $0.44 | $1.32 |
| Gemini 3.8 Flash | $1.50 | $7.50 |
| GLM-5.3 | $1.40 | $4.40 |
| GLM-5.3-Flash | $0.15 | $0.50 |
| Kimi K3 | $3.00 | $15.00 |
| Kimi K2.7 Code | $0.95 | $4.00 |
| Grok 4.6 | $2.00 | $6.00 |
Paid capabilities & connectors
These vendor-backed actions are charged only when an agent uses them. The provider shown is the service that actually fulfills the action.
Neural search
$0.01/run
Web content
$0.002/page
Web fetch
$0.004/page
Browser automation
$0.03/minute
Search results
$0.006/run
Image generation
$0.422/asset
Background removal
$0.01/asset
Video generation
$0.304/second
Image-to-video
$0.0488/second
Audio transcription
$0.0002/second
Connector call
$0.029/call
| Capability | Provider | Rate |
|---|---|---|
| Neural search | Exa | $0.01/run |
| Web content | Exa | $0.002/page |
| Web fetch | TinyFish | $0.004/page |
| Browser automation | TinyFish | $0.03/minute |
| Search results | DataForSEO | $0.006/run |
| Image generation | MiniMax Canvas | $0.422/asset |
| Background removal | Replicate | $0.01/asset |
| Video generation | BytePlus Seedance | $0.304/second |
| Image-to-video | Alibaba Cloud DashScope | $0.0488/second |
| Audio transcription | Fish Audio | $0.0002/second |
| Connector call | Pipedream | $0.029/call |
Compute
Billed per second while your agent's machine is awake. It suspends automatically when idle and wakes on the next message.
1 vCPU · 512 MB
1 vCPU · 1 GB
1 vCPU · 2 GB
2 vCPU · 512 MB
2 vCPU · 1 GB
2 vCPU · 2 GB
2 vCPU · 4 GB
4 vCPU · 1 GB
4 vCPU · 2 GB
4 vCPU · 4 GB
4 vCPU · 8 GB
8 vCPU · 2 GB
8 vCPU · 4 GB
8 vCPU · 8 GB
Suspended or parked
Storage
| Machine | Per hour awake | Per 24h awake | Always on, per month |
|---|---|---|---|
| 1 vCPU · 512 MB | $0.117 | $2.81 | $84.24 |
| 1 vCPU · 1 GB | $0.133 | $3.20 | $95.90 |
| 1 vCPU · 2 GB | $0.166 | $3.97 | $119.23 |
| 2 vCPU · 512 MB | $0.218 | $5.23 | $156.82 |
| 2 vCPU · 1 GB | $0.234 | $5.62 | $168.48 |
| 2 vCPU · 2 GB | $0.266 | $6.39 | $191.81 |
| 2 vCPU · 4 GB | $0.331 | $7.95 | $238.46 |
| 4 vCPU · 1 GB | $0.436 | $10.45 | $313.63 |
| 4 vCPU · 2 GB | $0.468 | $11.23 | $336.96 |
| 4 vCPU · 4 GB | $0.533 | $12.79 | $383.62 |
| 4 vCPU · 8 GB | $0.662 | $15.90 | $476.93 |
| 8 vCPU · 2 GB | $0.871 | $20.91 | $627.26 |
| 8 vCPU · 4 GB | $0.936 | $22.46 | $673.92 |
| 8 vCPU · 8 GB | $1.066 | $25.57 | $767.23 |
| Suspended or parked | $0 — idle agents suspend automatically; parked agents are not billed for compute | ||
| Storage | $0 — every agent includes 20 GB of persistent storage, kept even while parked, at no charge | ||
Estimate your monthly cost
$97.87
Frequently asked questions
- How does AgentSky's usage-based pricing work?
- You top up a prepaid credit balance and pay only for what your agent actually does — model tokens billed at published rates plus compute time metered per second while the agent is awake. There are no seats, no minimums, and no bill while an agent is parked.
- Can I use my existing Claude or ChatGPT subscription?
- Yes. Connect your Claude Pro or Max subscription and Claude Code agents run on it at $0 model usage. Connect your ChatGPT Plus or Pro plan and Codex agents run on it at $0 model usage. Compute is still metered per second.
- What counts as compute time?
- Compute is billed per second only while your agent's machine is awake and processing. Agents suspend automatically when idle and wake on the next message. Parked agents are never billed for compute.
- How much does storage cost?
- Storage is included at no charge. Every agent gets 20 GB of persistent storage, kept even while the agent is parked.
- Is the harness (Claude Code, Codex, Hermes) free?
- Yes. The agent harness you choose at launch — Claude Code, Codex, or Hermes — is free. You only pay for model tokens and compute.
Pay for exactly what you use
Top up a prepaid balance, launch an agent, inspect the itemized usage, and park it when you do not need active compute.