Tools
Interactive tools to help you decide, not just read. Everything is computed in your browser; nothing is sent.
AI cost: token API vs local VPS
Torn between a token-billed API (Mistral, OpenAI…) and self-hosting a model on a fixed-cost VPS? Estimate both based on your volume.
Break-even point : Above ~7,500 calls/month, the local VPS becomes cheaper than the API.
At this volume, the token-based API is cheaper (and faster, no server to manage).
Indicative estimate. Local cost excludes management time and CPU slowness; API cost varies by model and input/output ratio.
AI API price comparator
The price per million tokens (input / output) of the main AI APIs, side by side. Sort by input or output cost and filter by provider.
| Provider | Model | Input$/M tokens | Output$/M tokens | Context | |
|---|---|---|---|---|---|
| Mistral | Mistral Small | $0.2 | $0.6 | 128K | Pricing → |
| OpenAI | GPT-5 mini | $0.25 | $2 | 400K | Pricing → |
| DeepSeek | DeepSeek-V3 | $0.27 | $1.1 | 128K | Pricing → |
| Gemini 2.5 Flash | $0.3 | $2.5 | 1M | Pricing → | |
| Anthropic | Claude Haiku 4.5 | $1 | $5 | 200K | Pricing → |
| OpenAI | GPT-5 | $1.25 | $10 | 400K | Pricing → |
| Gemini 2.5 Pro | $1.25 | $10 | 1M | Pricing → | |
| Mistral | Mistral Large | $2 | $6 | 128K | Pricing → |
| Anthropic | Claude Sonnet 4.6 | $3 | $15 | 1M | Pricing → |
| Anthropic | Claude Opus 4.8 | $5 | $25 | 1M | Pricing → |
Prices in USD per 1M tokens (input / output). Indicative, last checked 2026-06-22 — they change often: always confirm on the provider's official page. Excludes discounts (cache, batch, volume).