Abliteration
Try Abliteration on us. Starter credit covers your first requests, then one key reaches all 4 Abliteration models, with the privacy posture shown on every route.
Abliterationabliteration
Chat, vision and reasoning models with cached-input pricing. Separate from Abliterate.ai.
These privacy labels describe Abliteration, the upstream model provider. Confidential requires all three: verified provider compute, end-to-end encryption and explicit zero data retention. Account and billing metadata are outside the content-retention claim.
| Provider | Abliteration |
|---|---|
| API operator country | Not verified Legal home of the API operator, not an inference-location guarantee. |
| Headquarters | Not verified |
| Routing status | Active |
| Provider website | https://abliteration.ai/ |
| Models | 4 public models |
| Credits routes | 4 |
| Zero data retention | yes All routes |
| Verified confidential inference | Not verified |
| Policy note | Core chat inference does not retain or train on prompts, completions or images. Operational and billing metadata are retained for the account lifetime and legal requirements. Optional web search/fetch are not ZDR and are not enabled by this integration. US inference is the default; EU residency requires a separate arrangement. No confidential-compute or renewable-energy claim has been verified. Policy source |
Documentation and policy sources
Inference locations
Provider API declarations are also available on each model's endpoint API. Direct chat and Responses replies separate advertised availability from serving location in usage.inference_location; the actual serving region remains unknown when not reported.
usage.provider_usage.gateway_region identifies the TrustedRouter gateway, not the provider's inference location. A value such as europe-west4 does not mean the model ran in Europe. usage.inference_location.serving_region remains null when the provider has not reported it. Older gateways may return the gateway-only field as provider_usage.region.
Where the upstream GPUs execute the model, not the provider's headquarters, API ingress, or data storage. ZDR and confidential inference do not by themselves restrict geography.
- Declared countries or regions
- United States
Provider-declared default inference location; EU residency is separately arranged. - Can locations change?
- Datacenter placement within the default region is not published.
- Provider region pinning
- EU residency requires contacting the provider; no documented per-request regional pin.
- Pinning through TrustedRouter
- No finer regional pin is supported by this integration.
- Infrastructure declaration
- United States by provider declaration, not an attested location.
- Evidence
- Published data-handling policy, default processing region.. Reviewed .
Provider selection alone is not a region pin. TrustedRouter's US provider filter selects company jurisdiction, not GPU location. A regional gateway hostname does not pin a downstream GPU. For a binding residency requirement, confirm the model, region, failover scope and enforced controls before sending sensitive data.
Measured performance
5 samplesContinuously sampled across Abliteration's routed models: p50 TTFT, effective throughput, and success rate. Effective throughput uses provider-reported output tokens over complete request time. Unsupported route and probe-configuration rows are separated from provider downtime. No prompt or output content stored.
| p50 TTFT | 1523 ms |
|---|---|
| Effective throughput | — |
| Uptime | 100.00% |
| Model | p50 TTFT | Effective throughput | Uptime | Config excluded | Availability samples |
|---|---|---|---|---|---|
| abliteration/abliterated-model-large-flash | 1269 ms | — | 100.00% | — | 1 |
| abliteration/abliterated-model-large-v2 | 1375 ms | — | 100.00% | — | 1 |
| abliteration/abliterated-model | 1523 ms | — | 100.00% | — | 1 |
| abliteration/abliterated-model-large | 1590 ms | — | 100.00% | — | 2 |
Abliteration performance history · Full provider & model leaderboard.
Models served by Abliteration.
Each row links to pricing, provider, benchmark, and API pages for the model.
| Model | AI IQ | Context | Input | Cached input | Output |
|---|---|---|---|---|---|
abliteration/abliterated-modelAbliterated Model |
— | 262,144 | $1.055/1M | $0.1055/1M | $3.165/1M |
abliteration/abliterated-model-largeAbliterated Large |
— | 1,000,000 | $3.165/1M | $0.3165/1M | $5.275/1M |
abliteration/abliterated-model-large-flashAbliterated Large Flash |
— | 1,048,576 | $1.055/1M | $0.1055/1M | $2.11/1M |
abliteration/abliterated-model-large-v2Abliterated Large v2 |
— | 1,000,000 | $3.165/1M | $0.3165/1M | $5.275/1M |
Questions
Does Abliteration have zero data retention?
TrustedRouter records Abliteration as supporting provider-level zero data retention based on the policy source linked on this page. This is a provider policy claim, separate from TrustedRouter's content-stateless real-time gateway and from end-to-end confidential compute.
Is Abliteration end-to-end encrypted?
TrustedRouter does not currently mark Abliteration as end-to-end encrypted at the provider boundary. The TrustedRouter gateway is still attested, but the selected provider normally receives the request in order to run the model. Use trustedrouter/e2e for the stronger route requirement.
Which Abliteration models are available through TrustedRouter?
This page currently lists 4 public Abliteration models, with live pricing, route count, context length, measured performance when available, and links to each model's provider and benchmark pages.