Kimi TokenHub

Menu
Your Cart
New: open-weight token packs are live, with per-million-token pricing from $0.50 per million tokens. Keys are delivered instantly, with no shipping. View packs →

open-weight token packs

Tokens, sold like software.

Prepaid packs for every open-weight family - Astra, Sol, Mini and Luna. Fixed token count, per-million price, no meter surprises.

Every open-weight pack is a virtual product. There is no shipping, no stock limit and no waiting. Buy 10+ units of a pack for 5% off, or 50+ units for 10% off. Prices are shown in your selected currency at checkout.

Open the full model directory →
01

Model directory

One directory for every open-weight models family.

open-weight, open-weight Sol, open-weight and open-weight, plus batch-rate and prompt-caching packs, all sold as token packs from one account with one checkout. Compare them here, then add the pack you need to the cart.

01

Compatibility

Bring your HTTP client

Every open-weight models answers on one endpoint - POST https://https://api.deepseek.com/v1 Drop in the key and the model id, there is no extra library to install.

02

Billing

Prepaid packs, no surprises

You pay for a fixed quota up front, so there is no monthly invoice, no auto-recharge and no overage.

03

Delivery

Virtual and instant

There is no shipping address and no delivery wait. The key appears in your account and inbox once the order completes.

04

Traceability

One key per order line

Every key is tied to an order line, a model and a quota, which makes spend easy to audit.

Fit check

No need to juggle billing accounts.

Tired of topping up balances and chasing invoices across projects? Buy the exact quota you need as a pack instead.

Buying credit directlyWith TokenHub packs
Accounts & keysCreate an organisation, set up billing and manage project keys yourself.Checkout once. A key for the model arrives with the order.
Budget controlWatch usage dashboards and set spending limits after the fact.Quota is fixed by the pack, so a key cannot overspend.
PaymentA card on file, billed monthly or topped up by auto-recharge.Pay per order with the methods offered at checkout.
DeliveryCredits applied to the whole organisation.One key per order line, with its own model and quota.

Compare current packs

Order + key trace

See how one order becomes a working key.

Your checkout stays the same. The pack you pick, the number of units and the order status decide how much quota the key carries and whether it is active.

StepSingle packStack unitsRefund
Orderthe model · 1 unitthe model · 12 unitsthe model · 1 unit
CheckoutNo shipping step10+ units −5%Complete
StatusCompleteCompleteRefunded
Keysk-••••••3f9aOne key per lineSuspended
Quota2M tokens24M tokensBack to Complete
Quota formulaunits purchased × pack size = quota on the key
Quota formula12 units × 2M tokens = 24M tokens
RuleA key is active only while its order is in a completed status

Pack prices already include the volume breaks for 10+ and 50+ units. Flagship packs currently carry up to 15% off list price. Check the model page for the live price in your currency.

First call

From cart to first request in three steps.

Registration is free and does not include model credit. Quota comes from the packs you buy.

  1. Create your account

    Register with an email. Keys are stored under your order history.

    Registration is free
  2. Buy a model pack

    Pick the model and the number of units, then pay at checkout. There is no shipping step.

    Volume breaks at 10+ and 50+ units
  3. Copy the key and call the model

    Once the order completes, copy the key from My Account › Order History and send your first request.

    Key issued automatically
Request
import os, requests

resp = requests.post(
    "https://https://api.deepseek.com/v1",
    headers={
        "Authorization": "Bearer " + os.environ["the platform_API_KEY"],
        "Content-Type": "application/json",
    },
    json={
        "model": "YOUR_MODEL_ID",
        "input": "Explain the CAP theorem in two sentences.",
    },
)
print(resp.json()["output"][0]["content"][0]["text"])

Quota is attached to the key, not to your account - a request never draws from a shared balance.

Pricing, without the guesswork

Buy a pack. Get the quota. Track every key.

Each model is sold as a fixed token pack, priced per million tokens at Hugging Face's own rate, and you can stack as many units as you need.

Pack unit
500K-200M tokens

Batch API packs are billed at half the standard rate, and prompt-caching packs bill cached input at 10% of the input price with 1.25x writes.

Volume breaks
10+ / 50+ units

5% off from 10 units and 10% off from 50 units. The discount is applied automatically.

Currency
Your selection

Prices convert to the currency picked in the header. You pay the amount shown at checkout.

Payment methods
Shown at checkout

Use any payment method the checkout page offers for your order.

Data handling

Know what we keep, and what we never see.

The store handles your order and the key it issues. The prompts and responses you send with that key never pass through the shop.

Orderkept

Order id, items and status

Keykept

Issued access key and its model id

Quotakept

Remaining balance attached to the key

Promptsnever

Never received by the storefront

Held with your orderNever reaches the storefront

Virtual delivery

One order. One key. Ready to call.

Token packs are virtual products. There is no warehouse and no courier involved. The store issues, suspends and restores access keys from the order status.

  1. Your orderModel pack × units

    Token packs never ask for a delivery address or a shipping method.

  2. TokenHubStatus · key · quota

    A key is generated for each virtual order line once payment completes.

  3. Your appHTTP client + key

    If an order leaves a completed status, its keys are suspended.

  4. Model endpointServes the request

    Keys, model ids and quota live under My Account › Order History.

TriggerOrder status: Complete
Key formatsk- (48 hex)
Also sent toOrder e-mail

Company

Who we are and how delivery works

Kimi TokenHub sells prepaid GPU-hour packs for open-weight models published on the Hugging Face Hub.

What we sell

Every product is one open-weight model plus a block of GPU hours: StarCoder2, Qwen2.5-Coder, MetaMath, Phi-2, BART, mBART, RWKV-5, NLLB-200, wav2vec2 XLSR, E5 Large, Jina v3 and Mamba.

How pricing works

T4 16GB at $0.50/hour, L4 24GB at $0.80/hour, A100 80GB at $2.50/hour. Fifty-hour packs carry a 10% volume discount.

How delivery works

The endpoint key appears in My Account / Order History once the order completes.

Support and refunds

Unused packs can be refunded within 14 days while no request has been made.

Ready to start?

Pick the model. Buy the pack. Get the key.

open-weight token packs from Hugging Face, delivered as virtual products.

Virtual deliveryNo shipping stepVolume breaks at 10 / 50 unitsKeys in your account
StarCoder2 7B · 10 hours on L4 - a prepaid inference-hour pack for bigcode/starcoder2-7b, ..
$8.00
Qwen2.5-Coder 7B · 10 hours on L4 - a prepaid inference-hour pack for Qwen/Qwen2.5-Coder-7..
$8.00
MetaMath 7B · 10 hours on L4 - a prepaid inference-hour pack for meta-math/MetaMath-7B-Ins..
$8.00
Qwen2.5-Coder 7B · 50 hours on L4 - a prepaid inference-hour pack for Qwen/Qwen2.5-Coder-7..
$36.00
Multilingual E5 Large · 10 hours on T4 - a prepaid inference-hour pack for intfloat/multil..
$5.00
mxbai Rerank Large · 10 hours on T4 - a prepaid inference-hour pack for mixedbread-ai/mxba..
$5.00
XTTS v2 voice cloning · 10 hours on T4 - a prepaid inference-hour pack for coqui/XTTS-v2, ..
$5.00
NLLB-200 distilled 600M · 10 hours on A100 - a prepaid inference-hour pack for facebook/nl..
$25.00
MetaMath 7B · 10 hours on L4 - a prepaid inference-hour pack for meta-math/MetaMath-7B-Ins..
$8.00
Qwen2.5-Coder 7B · 10 hours on L4 - a prepaid inference-hour pack for Qwen/Qwen2.5-Coder-7..
$8.00
Qwen2.5-Coder 7B · 50 hours on L4 - a prepaid inference-hour pack for Qwen/Qwen2.5-Coder-7..
$36.00
StarCoder2 7B · 10 hours on L4 - a prepaid inference-hour pack for bigcode/starcoder2-7b, ..
$8.00
Jina Embeddings v3 · 10 hours on T4 - a prepaid inference-hour pack for jinaai/jina-embedd..
$5.00
Multilingual E5 Large · 10 hours on T4 - a prepaid inference-hour pack for intfloat/multil..
$5.00
Multilingual E5 Large · 50 hours on T4 - a prepaid inference-hour pack for intfloat/multil..
$22.50
mxbai Rerank Large · 10 hours on T4 - a prepaid inference-hour pack for mixedbread-ai/mxba..
$5.00
wav2vec2 XLSR-53 · 10 hours on T4 - a prepaid inference-hour pack for facebook/wav2vec2-la..
$5.00
Whisper large-v3 · 10 hours on L4 - a prepaid inference-hour pack for openai/whisper-large..
$8.00
Whisper large-v3 · 50 hours on L4 - a prepaid inference-hour pack for openai/whisper-large..
$36.00
XTTS v2 voice cloning · 10 hours on T4 - a prepaid inference-hour pack for coqui/XTTS-v2, ..
$5.00
Mamba 2.8B · 10 hours on A100 - a prepaid inference-hour pack for state-spaces/mamba-2.8b,..
$25.00
NLLB-200 distilled 600M · 10 hours on A100 - a prepaid inference-hour pack for facebook/nl..
$25.00
NLLB-200 distilled 600M · 50 hours on A100 - a prepaid inference-hour pack for facebook/nl..
$112.50
Phi-2 · 10 hours on L4 - a prepaid inference-hour pack for microsoft/phi-2, billed at Hugg..
$8.00
RWKV-5 World 3B · 10 hours on L4 - a prepaid inference-hour pack for RWKV/rwkv-5-world-3b,..
$8.00
BLIP-2 OPT 2.7B · 10 hours on A100 - a prepaid inference-hour pack for Salesforce/blip2-op..
$25.00