Providers
Same weights. Different rules.
A model is only as uncensored as the host serving it. Some add classifiers, some add system prompts, some quantise until tool calls break. Here is what each one declares, and what we measure.
CSAM is prohibited on every provider, every model and every route through Open Uncensored. It is not a category we measure, and there is no setting that turns it off.
From each provider's public statements
Reported| Provider | Cyber | Bio | Chem | Weap | Drug | Adult | Viol | Hate | Self | OSINT | Fraud | Polit | CSAM |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Venice | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Blocked |
| Featherless | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Blocked |
| ArliAI | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Blocked |
| NanoGPT | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Blocked |
| OpenRouter | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Blocked |
| Abliteration.ai | Open | Open | Open | Open | Open | Open | Open | Open | Blocked | Open | Open | Open | Blocked |
| Chutes | n/t | n/t | n/t | n/t | n/t | n/t | n/t | n/t | n/t | n/t | n/t | n/t | Blocked |
| Atlas Cloud | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Blocked |
| Open Uncensored Cloud | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Open | Blocked |
Directory
9 hosts, one API.
Venice
Private inference with its own uncensored model line and a crypto native API.
- Billing
- Per token + plans
- Logging
- No logs
- Crypto
- Accepted
- Jurisdiction
- United States
- Uptime 30d
- 99.62%
- Models served
- 2
- Median speed
- 92 tok/s
- Median TTFT
- 256 ms
API from $0.20 in / $0.90 out per M tokens. App plans $18 to $200 a month.
Featherless
Flat rate serverless access to thousands of Hugging Face fine tunes, 781 of them tagged uncensored.
- Billing
- Per token + plans
- Logging
- No logs
- Crypto
- No
- Jurisdiction
- United States
- Uptime 30d
- 98.90%
- Models served
- 17
- Median speed
- 92 tok/s
- Median TTFT
- 260 ms
Basic $10 a month, Chat $25 a month unlimited tokens at 32K context. Developer plans from $50.
ArliAI
Unlimited token plans focused on creative writing and roleplay models.
- Billing
- Subscription
- Logging
- No logs
- Crypto
- No
- Jurisdiction
- Not published
- Uptime 30d
- 98.40%
- Models served
- 6
- Median speed
- 95 tok/s
- Median TTFT
- 240 ms
Monthly plans with unlimited tokens.
NanoGPT
Pay per prompt aggregator. No account needed, crypto checkout, TEE hosted models.
- Billing
- Per token
- Logging
- Transient
- Crypto
- Accepted
- Jurisdiction
- Not published
- Uptime 30d
- 99.10%
- Models served
- 10
- Median speed
- 75 tok/s
- Median TTFT
- 303 ms
Card from $1, BTC, Lightning and XMR from $0.10.
OpenRouter
General purpose router. Lists a handful of uncensored fine tunes alongside frontier models.
- Billing
- Per token
- Logging
- Transient
- Crypto
- Accepted
- Jurisdiction
- United States
- Uptime 30d
- 99.80%
- Models served
- 3
- Median speed
- 67 tok/s
- Median TTFT
- 321 ms
Pass through per token pricing plus a fee on credit purchases.
Abliteration.ai
Hosted abliterated frontier models, led by Abliterated Large v2 on GLM-5.3.
- Billing
- Per token + plans
- Logging
- No logs
- Crypto
- No
- Jurisdiction
- United States
- Uptime 30d
- 99.30%
- Models served
- 1
- Median speed
- 57 tok/s
- Median TTFT
- 277 ms
Around $3 in / $3 out per M tokens. Monthly tiers $20 to $200.
Chutes
Decentralised GPU network on Bittensor with very low per token prices.
- Billing
- Per token
- Logging
- Unknown
- Crypto
- Accepted
- Jurisdiction
- Decentralised
- Uptime 30d
- 96.70%
- Models served
- 5
- Median speed
- 68 tok/s
- Median TTFT
- 339 ms
Per token, often below cost. Many models served at int4.
Atlas Cloud
Image generation API serving open FLUX and Z-Image weights.
- Billing
- Per token
- Logging
- No logs
- Crypto
- No
- Jurisdiction
- Not published
- Uptime 30d
- 99.00%
- Models served
- 3
- Median speed
- n/a
- Median TTFT
- n/a
FLUX Schnell $0.003, Z-Image Turbo $0.01, FLUX Dev $0.012 per image.
Open Uncensored Cloud
Our own vLLM fleet for models nobody else serves yet.
- Billing
- Per token
- Logging
- No logs
- Crypto
- Accepted
- Jurisdiction
- United States
- Uptime 30d
- 99.20%
- Models served
- 7
- Median speed
- 71 tok/s
- Median TTFT
- 238 ms
At cost GPU pricing per M tokens.