Run Cu-Mai R1 70B without a GPU.
Roleplay by Steelskull, on Unrestrictable as cu-mai-r1-70b.
Cu-Mai R1 70B is Steelskull's roleplay model. It costs $1.7255 per million input tokens and $1.7255 per million output tokens, prepaid, with a 32,768-token context window. Call it at https://api.unrestrictable.ai/v1/chat/completions with model "cu-mai-r1-70b" and an Unrestrictable key, from any OpenAI SDK or client that takes a base URL.
Price per million tokens
- Input
- $1.7255
- Cached input
- $1.7255 (input rate)
- Output
- $1.7255
- Daily check
- Changed Nov 14
- As of
- 2026-11-16
First request
curl https://api.unrestrictable.ai/v1/chat/completions \ -H "Authorization: Bearer sk-unr-live-…" \ -H "Content-Type: application/json" \ -d '{"model": "cu-mai-r1-70b", "messages": [{"role": "user", "content": "Hello"}]}'Cu-Mai R1 70B settings for SillyTavern and Janitor AI
- SillyTavern API
- Chat Completion
- Source
- Custom (OpenAI-compatible)
- Custom endpoint
- https://api.unrestrictable.ai/v1
- Model
- cu-mai-r1-70b
- Context Size
- 32,768 tokens
- Janitor AI proxy URL
- https://api.unrestrictable.ai/v1/chat/completions
- Key prefix
- sk-unr-
Cu-Mai R1 70B specs
- Kind
- Roleplay
- Developer
- Steelskull
- Context window
- 32,768 tokens
- Input
- text
- Wire formats
- openai
Get a key
- Create a key in the console. Credit is prepaid from $10, with no subscription.
- Set OPENAI_BASE_URL to https://api.unrestrictable.ai/v1 and send your sk-unr- key as the bearer token.
- Send cu-mai-r1-70b as the model. The same key calls every model on Unrestrictable.
What Cu-Mai R1 70B is
- Base model
- Made by
- Steelskull
- Method
- SCE merge on a custom Hydroblated-R1 base
- License on the card
- llama3.3
- Downloads, 30 days to 2026-10-20
- 40
- Context window here
- 32,768 tokens
- Input
- text
From the SDK
from openai import OpenAI client = OpenAI(base_url="https://api.unrestrictable.ai/v1", api_key="sk-unr-live-…") completion = client.chat.completions.create( model="cu-mai-r1-70b", messages=[{"role": "user", "content": "Hello"}],)print(completion.choices[0].message.content)import OpenAI from 'openai'; const client = new OpenAI({ baseURL: 'https://api.unrestrictable.ai/v1', apiKey: 'sk-unr-live-…' }); const completion = await client.chat.completions.create({ model: 'cu-mai-r1-70b', messages: [{ role: 'user', content: 'Hello' }],});console.log(completion.choices[0].message.content);What Cu-Mai R1 70B is built to write
Cu-Mai R1 70B is a fiction fine-tune: trained on long-form stories and character dialogue, it holds a character's voice, follows a scenario and writes at length without the stock refusal layer.
Every request is routed to the build’s live route and settled to the token at the prices above. The daily real-model check sends a fixed prompt to every model and records whether the answer still matches; this model’s verdict today is “Changed Nov 14”.
Use it from the OpenAI SDK, from SillyTavern or Janitor, or from a coding agent. Keys and top-ups live in the console.
What Cu-Mai R1 70B costs, worked
A chat turn of 2,000 input and 500 output tokens at $1.7255 input and $1.7255 output per million tokens is $0.0043; 1,000 turns are $4.31. A long chat session of 100 replies of 300 tokens costs $0.742 at 4,000 tokens of context and $2.8126 at 16,000: every message resends the context, so size it to what the chat needs. All from prepaid credit, no subscription.
| Roleplay builds | Context | In / out per 1M | 1,000 chat turns |
|---|---|---|---|
| Manta Flash 1.0 | 16K | $0.07 / $0.56 | $0.42 |
| Manta Mini 1.0 | 8K | $0.07 / $0.56 | $0.42 |
| Gemma 4 12B Starry Ocean | 131K | $0.175 / $0.875 | $0.79 |
| Gemma 4 12B StationKeeper | 131K | $0.175 / $0.875 | $0.79 |
| Magidonia 24B v4.3 | 33K | $0.3535 / $0.4235 | $0.92 |
| MythoMax L2 13B | 4K | $0.404 / $0.404 | $1.01 |
| Cydonia 24B v2 | 33K | $0.4012 / $0.4828 | $1.04 |
| Gemma 4 31B Split-Untied | 262K | $0.35 / $1.05 | $1.23 |
| Cydonia 24B v4.3 | 131K | $0.484 / $0.60 | $1.27 |
| Manta Pro 1.0 | 66K | $0.21 / $1.75 | $1.30 |
| Artemis v1.1 | 262K | $0.35 / $1.575 | $1.49 |
| Gemma 4 31B Gembrain | 262K | $0.35 / $1.575 | $1.49 |
| Gemma 4 31B Isometry | 262K | $0.35 / $1.575 | $1.49 |
| Cydonia 24B v4.1 | 131K | $1.20 / $2.00 | $3.40 |
| Cu-Mai R1 70B | 33K | $1.7255 / $1.7255 | $4.31 |
| Hanami 70B x1 | 33K | $1.7255 / $1.7255 | $4.31 |
Cu-Mai R1 70B in SillyTavern and Janitor AI: what to set
The three fields are the base URL, the key and the model id, as in the settings above. What differs by build is the reply length, the reasoning switch and the context size, and those decide both whether a reply arrives whole and what each message costs.
- Context Size up to 32,768 tokens, the window the catalog lists for cu-mai-r1-70b. Every message resends the context, so set it to what the chat needs rather than the maximum.
- A fiction tune answers at once, with no reasoning step: a Max Response Length of 300 to 600 tokens fits a chat reply.
A request that fails before the model answers is not billed. A refusal written as ordinary text is billed like any answer, which is the reason to send this build rather than a stock model with a preamble: the build does not write one.
What Cu-Mai R1 70B costs by client
| Session | What is sent | At the live price |
|---|---|---|
| SillyTavern or Janitor AI, one evening | 100 replies of 300 tokens on 4,000 tokens of context | $0.742 |
| The same evening on a long chat | 100 replies of 300 tokens on 16,000 tokens of context | $2.8126 |
| A month of short turns | 1,000 turns of 2,000 input and 500 output tokens | $4.3138 |
Cu-Mai R1 70B questions
Cu-Mai R1 70B is a roleplay build from Steelskull, with a 32,768-token context window, taking text input. On Unrestrictable it answers as cu-mai-r1-70b.
No. Credit is prepaid from $10 and billed per token: input $1.7255 and output $1.7255 per million tokens, cached input $1.7255 (input rate).
About $0.0026 for 1,000 tokens in and 500 tokens out, at $1.7255 in and $1.7255 out per million tokens.
Up to 32,768 tokens, the context window the catalog lists for Cu-Mai R1 70B.
Yes. Send cu-mai-r1-70b as the model through the OpenAI SDK, or any client that takes a custom OpenAI base URL, with base URL https://api.unrestrictable.ai/v1.
In SillyTavern, set API to Chat Completion, Source to Custom (OpenAI-compatible) and the custom endpoint to https://api.unrestrictable.ai/v1, paste your key, then pick cu-mai-r1-70b from the list it loads. In Janitor AI, choose Proxy, set the URL to https://api.unrestrictable.ai/v1/chat/completions, paste your key and type cu-mai-r1-70b as the model. See the docs.
A custom Hydroblated-R1 base (TheSkullery’s L3.1x3.3-DS-Hydroblated-R1-70B v4.1), merged with SCE with EVA LLaMA 3.33, Euryale v2.3, Cirrus x1, Hanami x1, Anubis v1 and Negative_LLAMA, per its card.
The card lists the Llama 3.3 license (llama3.3).
Yes, the weights are on Hugging Face (Steelskull/L3.3-Cu-Mai-R1-70b, llama3.3, read 2026-10-20). Sending cu-mai-r1-70b to https://api.unrestrictable.ai/v1 needs no download and no GPU: 100 replies cost $0.742 at the live price.
$0.4314 for 100 chat turns of 2,000 input and 500 output tokens each, at $1.7255 input and $1.7255 output per million tokens, from prepaid credit with no subscription.
Yes, with the settings above: Chat Completion with a Custom (OpenAI-compatible) source and https://api.unrestrictable.ai/v1 as the endpoint in SillyTavern; the proxy URL https://api.unrestrictable.ai/v1/chat/completions and cu-mai-r1-70b as the model in Janitor AI. See the SillyTavern guide.