OpenAI · Anthropic · Google
One API key for every frontier model
Use GPT, Claude and Gemini from the SDKs you already have. Change the base URL, keep your code, and pay per token from a single balance.
from openai import OpenAI
client = OpenAI(base_url="https://mytenon.com/v1", api_key="sk-tx-...")
reply = client.chat.completions.create(
model="gpt-4o-mini",
messages=[{"role": "user", "content": "Hello!"}],
)
print(reply.choices[0].message.content)Run it as a batch and pay up to 25% less
Evals, labelling, summaries, translation, embedding backfills: anything that doesn't need an answer this second. Submit thousands of requests in one file and get the results within 24 hours, at a lower price per token.
See batch pricing| Model | Input / 1M | ||
|---|---|---|---|
| claude-fable-5 | $9.00 | −25% | |
| claude-opus-5-5 | $3.60 | −25% | |
| claude-sonnet-5-5 | $1.80 | −25% | |
| deepseek-v4-flash-0731 | $0.054 | −25% | |
Native APIs, not a lowest common denominator
Call OpenAI Chat & Responses, Anthropic Messages and Gemini through their own SDKs. Streaming, tools, prompt caching and thinking work as documented — and the OpenAI SDK can reach Claude and Gemini too.
Redundant capacity
Every model is served from several upstream accounts across vendors and cloud platforms. If one is rate-limited or down, your request moves to the next.
Pay only for tokens
No subscription and no minimum spend. Top up from $5 with a card, and see what every request cost — down to cached tokens.
Keys you control
Separate keys per project or teammate, each with an optional spending cap, rate limit and model allow-list. Disable a key instantly.
Works with the tools you use
Simple per-token pricing
| Model | Input / 1M | Output / 1M |
|---|---|---|
| gpt-oss-120b-ultra | $0.24 | $1.14 |
| claude-fable-5 | $12.00 | $60.00 |
| claude-opus-5-5 | $4.80 | $24.00 |
| claude-sonnet-5-5 | $2.40 | $12.00 |
| gemini-3.5-flash | $1.80 | $10.80 |
| gemini-3.7-flash | $0.90 | $4.50 |
Questions
Do I need to change my code?
Only the base URL and the API key. Requests and responses use each vendor's own format.
How is usage billed?
Per token, using the token counts the model provider reports — including cached and reasoning tokens — at the rates on the pricing page. Failed requests are free.
Does my credit expire?
No. Unused credit stays on your account.
Do you store my prompts?
No. We log metadata needed for billing — model, token counts, cost, latency — but not request or response content. Privacy