Nex-N2.5-Pro.
At 10,000,000 input + 2,000,000 output tokens per month, OpenRouter is the only listed host at $1.25/month.
Cheapest provider on the left.
Total monthly cost — input + output tokens combined.
Bill breakdown.
| Provider | Monthly total | |
|---|---|---|
| $1.25 | Sign up ↗ |
Rent the GPU instead of paying per token.
For an open-weights model like Nex-N2.5-Pro, you can rent a GPU and serve inference yourself. The math: cheapest GPU rental × 730 hours/month. Power is included in a rental's hourly price.
Assumes the GPU runs 24/7 at ~85% utilisation. If your traffic is bursty, you'll pay less for the API and probably more for the GPU (idle hours still cost rental). The breakeven analysis lives on the Self-host vs API breakeven tool.
Three steps to your monthly estimate.
-
01
Pick the model.
Use the search box to find an AI model — Claude, GPT, Llama, DeepSeek, Qwen, anything we track. The picker lists every model where at least one provider publishes per-token pricing.
-
02
Estimate volume.
Slide the monthly input + output token counts to match your expected workload. A typical chat app handles 1-10M input tokens per active user per month; an agent that re-reads context every turn can hit 100M+.
-
03
Read the spread.
The chart + table list every provider that hosts the model, sorted cheapest-first. Click a provider name to open its detail page — pricing history, throughput benchmarks, and the affiliate signup link.
Frequently asked.
How is the monthly bill calculated?
Total = (input rate × input tokens / 1M) + (output rate × output tokens / 1M), using each provider's published per-token prices from our latest reading.
Where does the pricing data come from?
From each inference provider's own pricing page or model-list API — Anthropic, OpenAI, OpenRouter, Together AI, Fireworks AI, DeepInfra, z.ai, Groq and others. We re-read each provider's model list every day and update its prices from that reading.
Should I always pick the cheapest provider?
Cheapest by $/M tokens isn't always cheapest in total. Watch for caching discounts an aggregator may not pass through, rate limits on smaller hosts that push you onto a pricier tier under load, and the extra hop an aggregator adds. For low-volume or bursty workloads the cheapest is usually right; for sustained production traffic, weigh throughput and reliability too.
What's the difference between OpenRouter and the model maker's direct API?
OpenRouter is an aggregator: it routes your request to one of several upstream providers. The model maker's direct API (e.g. api.anthropic.com for Claude) gives you its own price and native features such as prompt caching. Direct is usually cheaper at scale; an aggregator wins when you want one key for dozens of models.
Are input and output prices the same?
No — output tokens usually cost several times more than input, so the cheapest input provider isn't always the cheapest overall. The monthly bill column counts both; sort by it for the real total.
How does prompt caching affect this estimate?
Cached input tokens are usually much cheaper than fresh input. We don't model caching, because it depends on your workload: long system prompts reused across requests benefit, chat with fresh context each turn doesn't.
How often is this pricing refreshed?
Daily. We read each host's model list early each morning (UTC), so a provider's price change shows up within about a day.
What about fine-tuned variants?
Fine-tuned deployments are priced separately from the base model, often with a different structure (hourly compute plus per-token). This tool covers base-model token prices only; check the provider's own deployment pricing for fine-tuned costs.