OpenRouter Cost Estimator
Estimate OpenRouter API cost from a live model row, input and output tokens, monthly requests, and an optional platform fee. The result is a planning estimate for model usage, not an invoice.
Quick answer · Pricing data refreshed 2026-08-08 03:20:20
How does the OpenRouter cost estimator work?
Select the OpenRouter model row that matches your route, enter average input and output tokens per request, then add monthly request volume. The estimator multiplies the live token rates by that workload and lets you add a fee percentage when you want a broader credit-purchase estimate.
OpenRouter API cost calculator
Use average workload values from logs or a representative test. The calculator keeps model usage cost separate from hosting, taxes, exchange rates, and non-token features.
How to use the OpenRouter cost estimator
- Choose a real model row. Pick the same model and route you would use in production, not just the cheapest visible row.
- Measure the token mix. Use average input and output tokens from logs, a tokenizer estimate, or a representative benchmark.
- Project monthly traffic. Enter requests per month and include retries or fallback calls in the request count when they are part of your workflow.
- Add a fee only when verified. Keep the fee at zero for a model-token estimate, or enter a current account assumption to see a broader total.
- Validate before launch. Compare cost with latency, quality, context limits, availability, and the current official OpenRouter terms.
OpenRouter model pricing rows
These rows are loaded from the live AI Pricing Hub D1 pricing database and normalized to USD per 1M input and output tokens.
| Model | Provider | Input | Output | Blended | Context | Best fit |
|---|---|---|---|---|---|---|
| Auto Router audio,chat,image_gen,reasoning,tool_use,video,vision |
OpenRouter OpenRouter |
$-1000000.0000 | $-1000000.0000 | $-2000000.0000 | 2.0M | Reasoning and multi-step tasks |
| Auto Router (Beta) audio,chat,image_gen,reasoning,tool_use,video,vision |
OpenRouter OpenRouter |
$-1000000.0000 | $-1000000.0000 | $-2000000.0000 | 2.0M | Reasoning and multi-step tasks |
| Pareto Code Router chat,code |
OpenRouter OpenRouter |
$-1000000.0000 | $-1000000.0000 | $-2000000.0000 | 2.0M | Code and developer agents |
| Fusion chat |
OpenRouter OpenRouter |
$-1000000.0000 | $-1000000.0000 | $-2000000.0000 | 1.0M | Long-context analysis |
| Body Builder (beta) chat |
OpenRouter OpenRouter |
$-1000000.0000 | $-1000000.0000 | $-2000000.0000 | 128k | Long-context analysis |
| GTE-Base chat,tool_use |
Other DeepInfra |
$0.0050 | Free | $0.0050 | 512 | Chat, extraction, and classification |
| E5-Base-v2 chat,tool_use |
Other DeepInfra |
$0.0050 | Free | $0.0050 | 512 | Chat, extraction, and classification |
| paraphrase-MiniLM-L6-v2 chat,tool_use |
Other DeepInfra |
$0.0050 | Free | $0.0050 | 512 | Chat, extraction, and classification |
| all-MiniLM-L12-v2 chat,tool_use |
Other DeepInfra |
$0.0050 | Free | $0.0050 | 512 | Chat, extraction, and classification |
| bge-base-en-v1.5 chat,tool_use |
Other DeepInfra |
$0.0050 | Free | $0.0050 | 512 | Chat, extraction, and classification |
| multi-qa-mpnet-base-dot-v1 chat,tool_use |
Other DeepInfra |
$0.0050 | Free | $0.0050 | 512 | Chat, extraction, and classification |
| all-mpnet-base-v2 chat,tool_use |
Other DeepInfra |
$0.0050 | Free | $0.0050 | 512 | Chat, extraction, and classification |
| all-MiniLM-L6-v2 chat,tool_use |
Other DeepInfra |
$0.0050 | Free | $0.0050 | 512 | Chat, extraction, and classification |
| Qwen3 Embedding 8B chat,tool_use,code |
Alibaba Qwen DeepInfra |
$0.01 | Free | $0.01 | 32k | Code and developer agents |
| Qwen3 Embedding 8B chat,tool_use,code |
Alibaba Qwen Nebius |
$0.01 | Free | $0.01 | 32k | Code and developer agents |
| bge-m3 chat,tool_use |
Other DeepInfra |
$0.01 | Free | $0.01 | 8k | Chat, extraction, and classification |
| GTE-Large chat,tool_use |
Other DeepInfra |
$0.01 | Free | $0.01 | 512 | Chat, extraction, and classification |
| E5-Large-v2 chat,tool_use |
Other DeepInfra |
$0.01 | Free | $0.01 | 512 | Chat, extraction, and classification |
| Multilingual-E5-Large chat,tool_use |
Other DeepInfra |
$0.01 | Free | $0.01 | 512 | Chat, extraction, and classification |
| bge-large-en-v1.5 chat,tool_use |
Other DeepInfra |
$0.01 | Free | $0.01 | 512 | Chat, extraction, and classification |
| Qwen3 Embedding 4B chat,tool_use,code |
Alibaba Qwen DeepInfra |
$0.02 | Free | $0.02 | 33k | Code and developer agents |
| Text Embedding 3 Small chat,tool_use |
OpenAI OpenAI |
$0.02 | Free | $0.02 | 8k | Chat, extraction, and classification |
| LiquidAI/LFM2-8B-A1B chat,tool_use |
Other Liquid |
$0.01 | $0.02 | $0.03 | 33k | Chat, extraction, and classification |
| LiquidAI/LFM2-2.6B chat,tool_use |
Other Liquid |
$0.01 | $0.02 | $0.03 | 33k | Chat, extraction, and classification |
OpenRouter cost estimate examples
Examples show how workload shape changes the result; replace them with your own token distribution before budgeting.
Support chat
Try 2,000 input tokens, 800 output tokens, and 50,000 requests per month. This is a useful first pass for a short-response assistant, but it does not model tool calls or retries automatically.
Code or agent workflow
Keep the same request count, then raise output tokens and include fallback calls in monthly requests. The result makes output-heavy workloads visible for Qwen3 Embedding 8B before a production test.
What this OpenRouter estimate includes and excludes
Included
- Selected OpenRouter input and output token rates
- Per-request, per-1,000-request, and monthly model cost
- Optional fee percentage entered by the user
- Current model rows and context values available in the pricing database
Not included
- Hosting, observability, storage, taxes, or exchange rates
- Retries, fallback calls, tool calls, search, or image charges unless included in the request assumption
- Enterprise terms, credits, rate limits, or account-specific discounts
- Guaranteed invoice totals or future price changes
For production verification, review the official OpenRouter pricing page and official model catalog. The official pages are the source of truth when account terms or model prices change.
Related AI pricing tools
Move to a broader tool when the OpenRouter-only estimate is not enough for your decision.
OpenRouter cost estimator FAQ
It is a planning calculator that applies a selected OpenRouter model row to your average input tokens, output tokens, and monthly request volume. It shows model usage cost before any optional fee assumption.
It includes an optional fee percentage field rather than assuming that account terms are permanent. Leave it at zero for token cost only, or enter a current verified fee assumption before reviewing a broader total.
The page multiplies input tokens by the input rate, output tokens by the output rate, adds the two amounts, divides by one million, and multiplies by monthly requests. The optional fee is then applied to the model-cost subtotal.
A zero-priced row can produce a zero token-cost estimate in this table, but availability, rate limits, credits, queueing, and other account terms may still apply. Verify the current provider terms before relying on a free route.
Common causes include longer outputs, retries, fallback routing, tool or search calls, cache behavior, credits, fees, taxes, exchange rates, and price changes after the database refresh. Use production logs for a tighter forecast.