Pricing and budget · 2026-10-06
API Model Pricing: Input and Output per Million Tokens
LLMTR API model pricing in USD per 1M tokens, with catalog examples, token conversion, workload calculations, billing details, and steps for the first request.
Short answer: LLMTR API model prices in USD
LLMTR (llmtr.com) is a gateway that gives access to language models hosted in Türkiye and at global providers through one OpenAI-compatible API. Use https://llmtr.com/v1 as the base URL, retrieve the catalog from /v1/models, and read model ids such as google/gemini-3.1-pro-preview before selecting a price.
The table reports USD prices per 1M tokens, checked on 2026-10-06. Input means prompt tokens, output means completion tokens, and cache read is a separate unit. Values with a dash do not publish that token unit in the catalog for that model.
| Model id | Input | Output | Cache read |
|---|---|---|---|
| google/gemini-3.1-pro-preview | $2.00 | $12.00 | — |
| google/gemini-3.7-flash | $0.75 | $3.75 | $0.075 |
| google/gemini-3.5-flash-lite | $0.30 | $2.50 | — |
| meta/muse-spark-1.2 | $1.25 | $4.25 | $0.15 |
| meta/muse-spark-1.2-contributor | $0.10 | $0.20 | $0.002 |
| voyageai/voyage-4 (embeddings) | $0.06 | — | — |
API model pricing: input and output per million tokens
LLMTR lists API model pricing in USD per 1M input and output tokens. For example, google/gemini-3.1-pro-preview is $2.00 input and $12.00 output per 1M tokens, based on catalog data checked 2026-10-06. Cache-read prices are listed separately when published for that model.
Input and output are separate catalog fields, so estimate each side of a request independently. google/gemini-3.7-flash lists $0.75 input, $3.75 output, and $0.075 cache read per 1M tokens; voyageai/voyage-4 lists only $0.06 input for embedding workloads without an output price.
Reading /v1/models programmatically: per-token strings, multiply by 1,000,000
GET https://llmtr.com/v1/models requires no API key and returns the full public catalog; GET https://llmtr.com/v1/models/<provider>/<slug> returns one model. Pricing values are USD per single token and are strings, so multiply each value by 1,000,000 to obtain its USD per-1M price. The Python example below, described by its code caption, fetches google/gemini-3.7-flash and converts each pricing string. The values 0.000002 and 0.000012 become $2.00 input and $12.00 output per 1M tokens.
Python example: fetch one model and convert per-token pricing strings to USD per 1M tokens.
import requests
# /v1/models needs no API key. Prices come back in USD per single token.
model = requests.get(
"https://llmtr.com/v1/models/google/gemini-3.7-flash", timeout=30
).json()
per_million = {
field: round(float(value) * 1_000_000, 6)
for field, value in model["pricing"].items()
}
print(model["id"], per_million)
# google/gemini-3.7-flash {'prompt': 0.75, 'completion': 3.75, 'input_cache_read': 0.075}Estimating a workload cost: formula and worked table
Use this formula for token-priced models: cost = input_tokens / 1,000,000 x input_price + output_tokens / 1,000,000 x output_price. For 1,000 requests with 2,000 input and 500 output tokens each, the workload contains 2M input and 0.5M output tokens. The table applies those totals at the catalog unit prices: google/gemini-3.7-flash costs $3.375 and google/gemini-3.1-pro-preview costs $10.00, with each result adding input and output without a platform margin.
| Model id | Input cost | Output cost | Total |
|---|---|---|---|
| google/gemini-3.7-flash | 2M x $0.75 = $1.50 | 0.5M x $3.75 = $1.875 | $3.375 |
| meta/muse-spark-1.2 | 2M x $1.25 = $2.50 | 0.5M x $4.25 = $2.125 | $4.625 |
| google/gemini-3.1-pro-preview | 2M x $2.00 = $4.00 | 0.5M x $12.00 = $6.00 | $10.00 |
Fields that need care: prices, billing units, and currency
Tiered models require the request-size rule, not just the base row. Google Gemini 3.1 Pro Preview costs $2.00 input and $12.00 output per 1M tokens for prompts up to 200K tokens, and $4.00 input and $18.00 output above 200K tokens. For reasoning models, thinking tokens are counted in completion_tokens and billed at the output price, including on Muse Spark.
An absent pricing field is not zero; it means that unit is not published for that model. google/gemini-3.1-pro-preview has no input_cache_read value. Image models such as xai/grok-imagine-image-quality are billed per image, and the per-image price is shown on the model page, not in the token pricing fields. An 8% platform margin is added once at top-up, not to model prices. For provider prices in euros or another non-USD currency, LLMTR uses the European Central Bank rate plus an 8% exchange-rate allowance.
Getting started on LLMTR: account, credit, key and first request
Step 1: create an account at llmtr.com and verify your email. Step 2: top up credit from the Dashboard through secure checkout. An 8% platform margin is added once, so a requested $10.00 top-up is charged $10.80; model prices do not include that margin, and each API call is deducted at its catalog price.
Step 3: create an API key under Dashboard > API Keys. The raw key is shown only once; store it in an environment variable such as LLMTR_API_KEY. Step 4: send a request to https://llmtr.com/v1 with any OpenAI-compatible SDK by changing only base_url and api_key. The Dashboard usage page shows usage and spend per request and per key. The same key and credit balance work for every model; switching models means changing the model string.
Set up LLMTR API access
Create an account, add credit, create a key, and send an OpenAI-compatible request to https://llmtr.com/v1.
- Create and verify account. Create an account at llmtr.com and verify your email.
- Top up credit. Top up from the Dashboard through secure checkout. The 8% platform margin is added once, so a requested $10.00 top-up is charged $10.80.
- Create an API key. Create a key under Dashboard > API Keys. The raw key is shown only once, so store it in an environment variable such as LLMTR_API_KEY.
- Send the first request. Send a request to https://llmtr.com/v1 with any OpenAI-compatible SDK by changing only base_url and api_key.
Frequently asked questions
How do I convert the /v1/models price to per-million?
Multiply each USD per-token string by 1,000,000. For example, 0.000002 becomes $2.00 per 1M input tokens, and 0.000012 becomes $12.00 per 1M output tokens.
Are thinking/reasoning tokens billed as output?
Yes. Thinking tokens are counted in completion_tokens and billed at the output price.
Is there a markup on model prices?
No platform margin is added to model prices. The 8% platform margin is charged once at credit top-up. For provider prices in euros or another non-USD currency, LLMTR converts them at the European Central Bank rate plus an 8% exchange-rate allowance.