The API for uncensored open models.

Curated open-weight models served without a censorship layer — they answer as published, with no provider-side moderation rewriting your results. OpenAI-compatible, pay per token.

Security research

Probe real attack prose without a filter rewriting the finding.

Red-teaming & evals

Build eval sets against a model that answers the test case instead of refusing it.

Fiction & long-form

131K of context holds a whole manuscript — crime, horror, and adult themes included.

Agents that don’t stall

Tool calls and reasoning verified before launch, so a loop never dies on a refusal.

unbound/qwen3.8-27b131K context · ~92 tok/s · one model today, more in verificationView catalog →
The model

One model today. Deliberately.

We add a route only after it passes verification, so the catalog grows slowly and nothing in it is a surprise.

unbound/qwen3.8-27b

Qwen3.8 27B Uncensored

A community build of Qwen3.8 27B with the refusal layer removed — nothing else changed. Served in FP8 with a 131K context window, with the exact checkpoint named in every response.

weights · orcarouter/Qwen3.8-27B-Uncensored-FP8 ↗

Context131,072
Decode~92 tok/s
Input$0.20 / M
Output$7.00 / M
Tool callsverified
Reasoningverified
Structured outputverified
Moderationnone
Compatibility

Your code already works.

Any OpenAI SDK, any framework, any agent harness — the entire migration is one line.

  1. 1Sign in with an email code and create an API key.
  2. 2Point base_url at our endpoint.
  3. 3Top up a prepaid balance and pay per token. That’s all.
client = OpenAI(
- base_url="https://api.openai.com/v1",
+ base_url="https://api.unboundmodels.com/v1",
 api_key=os.environ["UNBOUND_API_KEY"],
)
What you can check

Every promise here has something behind it.

Verified before it ships

A model reaches the catalog only after it runs the full harness suite. If something fails, we don’t advertise it.

  • tool calling · multi-turnpass
  • reasoning parserpass
  • structured output · JSON schemapass
  • 131K retrievalpass

You know what answered

Every response names the exact checkpoint that produced it. No swaps, no hidden system prompt in front of yours.

"model": "unbound/qwen3.8-27b",
"unbound": {
  "checkpoint": "orcarouter/Qwen3.8-27B-Uncensored-FP8",
  "revision": "a41c9f2",
  "system_prompt": none
}

Nothing of yours is kept

Prompts and completions are never written to disk. Here is the full list of what we do keep.

We store

  • token counts
  • timestamps
  • model and key id

We never store

  • your prompts
  • completions
  • tool arguments
How serving works

The model sleeps. Your first request wakes it.

Dedicated GPUs are expensive to keep idle, so we don’t — and we don’t hide it either. When a model is cold, the stream opens immediately and tells you what is happening, so your client stays connected instead of timing out.

Warm-up is never billed: the meter starts with the first generated token. Once warm, requests stream at ~92 tok/s with a 95%+ prefix-cache hit rate on agent workloads.

live request · streaming
$ curl api.unboundmodels.com/v1/chat/completions \
 -d '{"model":"unbound/qwen3.8-27b", "stream":true}'

event: warming model waking · 38s · not billed
event: ready 131,072 ctx · tools · reasoning

data: {"delta":{"content":"Sure — here is the"}}
data: {"delta":{"content":" full breakdown"}}
Why here

Where else can you run this model?

We went looking. Every option we found, ours included.

OptionHas an uncensored buildContextWorks with agentsPrice
OpenRouter, Together,DeepInfra, Fireworks, Novitano — official checkpoints only
Wiro AIhosts oneyes262Kown run APInot published
Featherlesscommunity re-uploadsten of them32Kyes$25–50 / mo
Self-hostingweights from Hugging Faceyes262Kyou build ita GPU
UnboundModelsone verified checkpointyes131Kyes — tools verifiedper token, from $10

Checked against public catalogs in August 2026. Our own catch: the model sleeps, so a cold request waits ~40 s — shown in the stream, never billed.

Pricing

Pay as you go. Nothing else.

No subscription, no idle fees, no minimum. A prepaid balance that never expires, and no bill for the time the model spends waking up.

Input
$0.20 / M
Cached input
$0.05 / M
Output
$7.00 / M

Top up from $10. Cached input applies automatically on repeated prefixes, which is most of an agent loop.

Example · two-hour agent session, 40 tool-calling turns
input · 1.4 M$0.28
cached input · 12.6 M$0.63
output · 96 K$0.67
warm-up · 3 wakes$0.00
session total$1.58

The example is a worked calculation at the rates above, not a measured session — your numbers depend on how much context you reuse.

Straight answers

The questions you were going to ask anyway.

Is this legal?

Yes. These are published open-weight models, served under their own licenses. We are a hosting provider with published terms, operating from the EU.

So there are no rules at all?

There is no moderation layer between you and the model. There is still an acceptable-use policy: illegal use gets an account closed. Removing a filter does not remove the law.

Do you read or keep my prompts?

No. Prompts and completions are never stored, and nothing you send is used for training — ours or anyone else’s.

What happens when my balance runs out?

Requests stop until you top up. No overdraft, no invoice, no auto-charge to remember to cancel.

Why does the first request take 40 seconds?

The GPU sleeps when nobody is using it — that is why the price is what it is. The stream tells you it is waking, and the wait is never billed.

When do more models arrive?

As fast as verification allows. Tell us what you want hosted — requests shape the order we test in.