OpenAI
Try OpenAI on us. Starter credit covers your first requests, then one key reaches all 29 OpenAI models, with the privacy posture shown on every route.
OpenAIopenai
These privacy labels describe OpenAI, 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 | OpenAI |
|---|---|
| API operator country | US Legal home of the API operator, not an inference-location guarantee. |
| Headquarters | Not verified |
| Routing status | Active |
| Provider website | https://openai.com/ |
| Models | 29 public models |
| Credits routes | 29 |
| Zero data retention | yes Reviewed TR-funded routes only; check each model |
| Verified confidential inference | Not verified |
| Policy note | Contracted Zero Data Retention is active for TrustedRouter's managed OpenAI account, effective July 28, 2026, and live enforcement was verified on July 29, 2026. This guarantee applies only to reviewed TrustedRouter-funded routes; customer BYOK credentials use the data controls on the customer's own OpenAI organization or project. The organization and Default project were rechecked on October 28, 2026: ordinary ZDR, not Private Safety Processing (PSP). PSP retains encrypted safety records and never qualifies for TrustedRouter's strict ZDR filter. Models with persistent cache state are also excluded from strict ZDR. New models do not automatically inherit ZDR approval. Policy source |
| Strict ZDR and safety retention | Ordinary ZDR is active on the reviewed managed project (checked October 28, 2026). PSP is not enabled there. Future models need their own review. Models with persistent cache state are excluded from strict ZDR even on this ordinary ZDR account. If PSP is enabled, encrypted safety records remain in TrustedRouter-controlled storage for at least 30 days; those routes must be labeled PSP and excluded from strict ZDR. BYOK never inherits our project settings. Use provider.min_privacy="zdr" to reject retained or unverified routes, including during fallback. |
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
- Not verified
No inference-location declaration has been verified for this provider. - Can locations change?
- Not verified. Do not assume that the serving location is fixed.
- Provider region pinning
- Not verified.
- Pinning through TrustedRouter
- No verified provider-specific location pin documented here.
- Infrastructure declaration
- Inference location not verified; no country-specific residency commitment.
- Evidence
- Not yet 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
499 samplesContinuously sampled across OpenAI'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 | 1221 ms |
|---|---|
| Effective throughput | — |
| Uptime | 100.00% |
| Model | p50 TTFT | Effective throughput | Uptime | Config excluded | Availability samples |
|---|---|---|---|---|---|
| openai/gpt-4.1 | 746 ms | — | 100.00% | — | 2 |
| openai/gpt-4o-mini | 772 ms | — | 100.00% | — | 3 |
| openai/gpt-4.1-mini | 803 ms | — | 100.00% | — | 2 |
| openai/gpt-6-luna | 1220 ms | — | 100.00% | — | 6 |
| openai/gpt-4o | 1221 ms | — | 100.00% | — | 3 |
| openai/gpt-5.4-mini | 1363 ms | — | 100.00% | — | 3 |
| openai/gpt-6-sol | 1376 ms | — | 100.00% | — | 2 |
| openai/gpt-5.6-sol | 1528 ms | — | 100.00% | — | 1 |
| openai/gpt-5.1 | 1613 ms | — | 100.00% | — | 2 |
| openai/gpt-6-astra | 1859 ms | — | 100.00% | — | 1 |
| openai/gpt-5.2 | 2250 ms | — | 100.00% | — | 1 |
| openai/gpt-5-mini | 2600 ms | — | 100.00% | — | 1 |
| openai/gpt-5 | 3156 ms | — | 100.00% | — | 1 |
| openai/text-embedding-3-large | — | — | 100.00% | — | 467 |
| openai/text-embedding-3-small | — | — | 100.00% | — | 4 |
OpenAI performance history · Full provider & model leaderboard.
Models served by OpenAI.
Each row links to pricing, provider, benchmark, and API pages for the model.
| Model | AI IQ | Context | Input | Cached input | Output |
|---|---|---|---|---|---|
openai/chat-latestchat-latest Retention unverified - not strict ZDR
|
— | 0 | $5.275/1M | $0.5275/1M | $31.65/1M |
openai/gpt-3.5-turboOpenAI: GPT-3.5 Turbo ZDR - TR-funded
|
— | 16,385 | $0.5275/1M | Not published | $1.5825/1M |
openai/gpt-4.1OpenAI: GPT-4.1 ZDR - TR-funded
|
— | 1,047,576 | $2.11/1M | $0.5275/1M | $8.44/1M |
openai/gpt-4.1-miniOpenAI: GPT-4.1 Mini ZDR - TR-funded
|
— | 1,047,576 | $0.422/1M | $0.1055/1M | $1.688/1M |
openai/gpt-4oOpenAI: GPT-4o ZDR - TR-funded
|
IQ 85#168 | 128,000 | $2.6375/1M | $1.31875/1M | $10.55/1M |
openai/gpt-4o-miniOpenAI: GPT-4o-mini ZDR - TR-funded
|
— | 128,000 | $0.15825/1M | $0.079125/1M | $0.633/1M |
openai/gpt-5OpenAI: GPT-5 ZDR - TR-funded
|
IQ 111#73 | 400,000 | $1.31875/1M | $0.131875/1M | $10.55/1M |
openai/gpt-5-miniOpenAI: GPT-5 Mini ZDR - TR-funded
|
— | 400,000 | $0.26375/1M | $0.026375/1M | $2.11/1M |
openai/gpt-5-nanoOpenAI: GPT-5 Nano ZDR - TR-funded
|
IQ 95#146 | 400,000 | $0.05275/1M | $0.01/1M | $0.422/1M |
openai/gpt-5.1OpenAI: GPT-5.1 ZDR - TR-funded
|
IQ 111#74 | 400,000 | $1.31875/1M | $0.131875/1M | $10.55/1M |
openai/gpt-5.2OpenAI: GPT-5.2 ZDR - TR-funded
|
IQ 118#54 | 400,000 | $1.84625/1M | $0.184625/1M | $14.77/1M |
openai/gpt-5.4-miniOpenAI: GPT-5.4 Mini ZDR - TR-funded
|
IQ 111#75 | 400,000 | $0.79125/1M | $0.079125/1M | $4.7475/1M |
openai/gpt-5.5OpenAI: GPT-5.5 Encrypted cache - up to 24 hours
|
IQ 129#18 | 1,050,000 | $5.275/1M | $0.5275/1M to $1.055/1M | $31.65/1M |
openai/gpt-5.6-lunaOpenAI: GPT-5.6 Luna Encrypted cache - up to 24 hours
|
IQ 124#31 | 1,050,000 | $0.211/1M | $0.0211/1M | $1.266/1M |
openai/gpt-5.6-solOpenAI: GPT-5.6 Sol Encrypted cache - up to 24 hours
|
IQ 133#13 | 1,050,000 | $4.22/1M | $0.422/1M to $0.844/1M | $21.1/1M |
openai/gpt-5.6-terraOpenAI: GPT-5.6 Terra Encrypted cache - up to 24 hours
|
IQ 130#15 | 1,050,000 | $2.11/1M | $0.211/1M | $12.66/1M |
openai/gpt-6-astraOpenAI: GPT-6 Astra Encrypted cache - up to 24 hours
|
IQ 140#4 | 1,050,000 | $10.55/1M | $1.055/1M to $2.11/1M | $52.75/1M |
openai/gpt-6-lunaOpenAI: GPT-6 Luna Encrypted cache - up to 24 hours
|
IQ 125#27 | 1,050,000 | $0.1055/1M | $0.01055/1M to $0.0211/1M | $0.5275/1M |
openai/gpt-6-solOpenAI: GPT-6 Sol Encrypted cache - up to 24 hours
|
IQ 137#7 | 1,050,000 | $2.11/1M | $0.211/1M | $10.55/1M |
openai/gpt-6.1-solOpenAI: GPT-6.1 Sol Encrypted cache - up to 24 hours
|
IQ 141#1 | 1,050,000 | $2.11/1M | $0.1055/1M to $0.211/1M | $10.55/1M |
openai/gpt-image-2.5-flareGPT Image 2.5 Flare ZDR - TR-funded
|
— | 0 | $5.275/1M | $1.31875/1M | $31.65/1M |
openai/gpt-image-2.5-sunburstGPT Image 2.5 Sunburst ZDR - TR-funded
|
— | 0 | $5.275/1M | $1.31875/1M | $31.65/1M |
openai/o3OpenAI: o3 ZDR - TR-funded
|
IQ 103#112 | 200,000 | $2.11/1M | $0.5275/1M | $8.44/1M |
openai/sora-2OpenAI Sora 2 Retention unverified - not strict ZDR
|
— | 2,500 | selected route | Not published | selected route |
openai/sora-2-proOpenAI Sora 2 Pro Retention unverified - not strict ZDR
|
— | 2,500 | selected route | Not published | selected route |
openai/text-embedding-3-largeOpenAI Text Embedding 3 Large ZDR - TR-funded
|
— | 8,191 | $0.13715/1M | Not published | selected route |
openai/text-embedding-3-smallOpenAI Text Embedding 3 Small ZDR - TR-funded
|
— | 8,191 | $0.0211/1M | Not published | selected route |
openai/text-embedding-ada-002OpenAI Text Embedding Ada 002 ZDR - TR-funded
|
— | 8,191 | $0.1055/1M | Not published | selected route |
openai/whisper-1OpenAI Whisper Retention unverified - not strict ZDR
|
— | 7 MiB, 10 minutes | $0.00633/audio minute | Not applicable | Included |
Questions
Does OpenAI have zero data retention?
Reviewed managed OpenAI routes can qualify for zero data retention. Check the exact model's endpoint policy: safety-retention, durable-cache and unreviewed routes are excluded. BYOK does not inherit this classification. Use provider.min_privacy=zdr to enforce eligibility.
Is OpenAI end-to-end encrypted?
TrustedRouter does not currently mark OpenAI 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 OpenAI models are available through TrustedRouter?
This page currently lists 29 public OpenAI models, with live pricing, route count, context length, measured performance when available, and links to each model's provider and benchmark pages.