Anton pricing
Start with open-source Anton, use hosted Anton on Minds Hub, and choose whether to bring your own model keys or use the Minds Hub model router.
Plans
The Anton subscription is separate from LLM usage. In either setup, you can bring your own model keys or use the Minds Hub model router and pay the token rates below.
$0/ month
Run Anton yourself. Use your own keys or the Minds Hub router.
Best when you want local control, want to inspect or modify the code, or want to choose exactly how Anton connects to models.
$9.95/ month
First 7 days free.
Best when you want hosted onboarding, token visibility, and a simpler managed setup.
Important: Anton users can select their own models in both local and hosted setups: bring your own API keys, run local models where supported, or use the Minds Hub model router according to the token prices on this page.
Model choice
Anton can work with frontier models, popular open-source and open-weight models, or your own model endpoints.
Pay for usage directly with the provider, and Anton uses your selected model.
Use a single router for multiple model families. Minds Hub bills usage with a simple per-token rate card.
Use stronger models for planning and reasoning, and cheaper models for lighter summarization or formatting work.
LLM token pricing
Prices below are per 1 million tokens in USD. They apply when users choose the Minds Hub model router instead of bringing their own model keys.
| Model | Input / 1M | Output / 1M | Best for |
|---|---|---|---|
| Claude Opus 4.7 | $5.25 | $26.25 | Hard reasoning, complex analysis, long context |
| Claude Sonnet 4.6 | $3.15 | $15.75 | Everyday agentic work, SQL, coding, analysis |
| Claude Haiku 4.5 | $1.05 | $5.25 | Fast classification, extraction, light drafting |
| GPT-5.5 | $5.25 | $31.50 | High-capability agentic work and coding |
| GPT-5.4 | $2.63 | $15.75 | Balanced reasoning, speed, and cost |
| GPT-5.4 mini | $0.79 | $4.73 | Lower-cost agent steps and high-volume tasks |
| Model | Input / 1M | Output / 1M | Best for |
|---|---|---|---|
| Kimi K2.6 | $1.00 | $4.20 | Strong open-model reasoning and agent tasks |
| DeepSeek V4 Pro | $1.83 | $3.65 | Reasoning-heavy work at lower output cost |
| DeepSeek V3 family | $0.59 | $1.76 | Cost-efficient analysis and coding workflows |
| GLM 5.1 | $1.47 | $4.62 | General-purpose open-model tasks |
| GPT OSS 120B | $0.16 | $0.63 | Very low-cost open-weight routing |
The table shows standard token rates. When a model supports caching, applicable savings are handled automatically. Model availability and token prices may change.
FAQ
No. You can bring your own model API keys and pay for usage directly. Minds Hub model pricing only applies when you use the Minds Hub router.
It covers hosted Anton access through Minds Hub. LLM usage is billed separately when you use the Minds Hub model router.
Anton can use different models for different parts of a task. Separating the product subscription from token usage keeps costs transparent.
Yes. Anton is designed for model choice. You can use frontier models, open models, or your own model keys depending on the task.
Get started
Run it locally from GitHub, or start with Minds Hub and use the model router when you want hosted convenience.
