Choose your workload
Models for extracting, summarizing and reasoning over document content.
Compare the recorded score and API price.
For native PDF input, contact Roten to confirm the request format. The examples here send document text.
Consider the trade-off between cost and the recorded score.
Document summaries and reasoning using extracted text.
Prices apply to the stated model settings. Token input rates cover uncached input; output tokens are priced separately. Cache usage and image or video settings can change the charge.
Input formats, model capabilities and prices.
| Reference | Claude Fable 5 HighPerformance candidate | Kimi K2.6Cost candidate |
|---|---|---|
| Input → output | text, image, document → text | text → text Documents use extracted text, not native PDF uploads. |
| Uncached input / 1M tokens | $7.00 | $0.665 |
| Output / 1M tokens | $35.00 | $2.80 |
| Cost comparison | Scoped price reference | 90.5% lower input cost · 92% lower output cost |
| Reference Arena score | 1496 | 1451 |
| Billing details | Uncached input and output are priced separately per million tokens. | Uncached input and output are priced separately per million tokens. |
Implementation
Copy an available model request, replace the API key with your own, and run.
API model ID: kimi-k2.6
Document summaries and reasoning using extracted text. Native PDF uploads are not supported by these examples.
Documents use extracted text, not native PDF uploads.
This example sends extracted or sample document text. Input format: text.
Copy the Python or cURL test for Kimi K2.6. Replace the API-key placeholder with your own key and run.
Document text test · The sample document is included in the request as text.
import requests
response = requests.post(
url="https://api.roten-ai.com/v1/chat/completions",
headers={
"Authorization": "Bearer YOUR_ROTEN_API_KEY",
},
json={
"model": "kimi-k2.6",
"messages": [
{
"role": "user",
"content": "Summarize this sample project memo in two bullets.\n\nThe product team is evaluating an AI API for customer-support replies. The trial will compare response usefulness and request cost. Engineering will record the results before selecting a model."
}
],
"max_tokens": 256,
"thinking": {
"type": "disabled"
},
"stream": False
},
timeout=180,
)
print(response.json())
response.raise_for_status()curl --fail-with-body 'https://api.roten-ai.com/v1/chat/completions' \
-H 'Content-Type: application/json' \
-H 'Authorization: Bearer YOUR_ROTEN_API_KEY' \
-d '{
"model": "kimi-k2.6",
"messages": [
{
"role": "user",
"content": "Summarize this sample project memo in two bullets.\n\nThe product team is evaluating an AI API for customer-support replies. The trial will compare response usefulness and request cost. Engineering will record the results before selecting a model."
}
],
"max_tokens": 256,
"thinking": {
"type": "disabled"
},
"stream": false
}'Python requires requests: install once with python3 -m pip install requests. cURL examples run directly in your terminal.
Evidence and limits
Reference Arena scores, API tests and billing rates describe different aspects of each candidate.
Built for production
Talk to us about volume pricing, model selection and migration.