Find a model

Browse the full catalog →

OpenAI compatible API · Attested · Public status

US AI Models & US-Operated Providers

Find open-weight AI models from US labs and compare US-operated providers. See the Liberty shortlist and routing controls; operator country is not data residency.

Verify gateway
Onebase URL to migrate
100sof models and routes
0prompt or output logs. Always.

Start with open weights. Choose the operator separately.

Looking for American open-weight AI? Start with the US-lab shortlist below, then compare the serving provider's price, policy and measured performance. Liberty combines US-lab open-weight components behind one model ID; individual models remain available for direct evaluation.

Three separate checks: the lab that built the model, the company operating the endpoint, and the physical inference location. A US company is not a data-residency guarantee. Use the request controls below for operator jurisdiction and verify processing-location requirements separately.

American labs, and the operators who serve them

"US AI models" is two questions. This page answers both, separately.

Who built it and where your prompt goes are different facts. 193 models here come from 15 US labs: OpenAI, Google DeepMind, Anthropic, Meta, NVIDIA, xAI, Recraft, Microsoft Research, ElevenLabs, ScaleDown, Reflection AI, Thinking Machines Lab, IBM Research, Inception Labs, TrustedRouter. 49 of 116 providers are US-headquartered.

Need US-only routing? Set provider.jurisdiction to us. It fails closed when nothing qualifies.

Browse US-provider routes Review provider policies

us_jurisdiction.pyPython
from openai import OpenAI

client = OpenAI(
    base_url="https://api.trustedrouter.com/v1",
    api_key="sk-tr-...",  # your TrustedRouter key
)

reply = client.chat.completions.create(
    model="trustedrouter/liberty-4.1",
    messages=[{"role": "user",
               "content": "Draft the incident summary."}],
    extra_body={"provider": {"jurisdiction": "us"}},
)

Liberty: US labs, open weights, one route

open weight

Our own family. Each Liberty id calls several component models and returns one answer — and every component is an open-weight model from a US lab. Both facts are computed from the catalog when this page renders, so if a component were ever swapped the counts would move and the claim would retire itself.

Route Context Component models From US labs US-operated route available
trustedrouter/liberty-4.1
TrustedRouter Liberty 4.1
1,048,576 4 of 4 yes
trustedrouter/liberty-4.0
TrustedRouter Liberty 4.0
1,048,576 3 of 3 yes
trustedrouter/liberty-3.0
TrustedRouter Liberty 3.0
1,048,576 5 of 5 yes
trustedrouter/liberty-2.0
TrustedRouter Liberty 2.0
262,144 5 models once the nested Liberty presets are expanded 5 of 5 yes
parasail/liberty-2.0
Parasail Liberty 2.0
262,144 5 models once the nested Liberty presets are expanded 5 of 5 yes
trustedrouter/liberty-1.0-1m
TrustedRouter Liberty 1.0 1M
1,048,576 1 of 1 yes
trustedrouter/liberty-1.0
TrustedRouter Liberty 1.0
262,144 3 of 3 yes

Two limits worth stating plainly. First, the component models are open-weights models that many providers serve, and Liberty's default candidate pool includes operators outside the US; set provider.jurisdiction to us when the serving provider has to be US-operated. Second, operator jurisdiction is not datacentre location — a US-registered company can run inference on hardware elsewhere, and the catalog records the company, not the machine.

Measurement for these routes is published rather than summarised here: our own DRACO run for Liberty 2.0 is written up in the Liberty 2.0 post, and live route behaviour is on the leaderboard.

Open weights from US labs

55 routes

The labs themselves. Weights you can download, run anywhere, and inspect.

Browse all open-weight routes

Live catalog, as of November 2026

Two separate facts: who operates the endpoint, and who built the model.

Counts below are read from the routes configured in TrustedRouter right now. Operator country is the legal home of the company running the endpoint a request reaches, taken from that company's own terms, privacy policy, or regulatory filing. It is not a claim about which datacentre serves the request.

49providers operated from the United States
15labs with models here
193models from those labs
116providers in the catalog
564callable catalog models
36providers whose operator we could not establish

Providers operated from the United States

49 of 116

Send a request to one of these and the operator is a company headquartered in the United States — head office, not place of registration. Privacy tier is a separate column: it is the highest bar we have recorded from that provider's published policy.

Provider Operator country Privacy tier Recorded claims Models served Policy source
DeepInfradeepinfra United States No-store no verified privacy claim 90 policy
Novita AInovita United States Standard no verified privacy claim 90 policy
Atlas Cloudatlas-cloud United States Standard no verified privacy claim 69 policy
GMI Cloudgmi United States Standard no verified privacy claim 68 policy
Venicevenice United States Standard no verified privacy claim 65 policy
Featherless AIfeatherless United States Zero retention ZDR 60 policy
IO Intelligenceio-net United States Standard no verified privacy claim 40 policy
OpenAIopenai United States Standard ZDR on prepaid 40 policy
Amazon Bedrockamazon-bedrock United States Standard no verified privacy claim 39 policy
Cloudflare Workers AIcloudflare-workers-ai United States Zero retention ZDR 27 policy
Parasailparasail United States Zero retention ZDR 25 policy
Weights & Biaseswandb United States Standard no verified privacy claim 22 policy
Basetenbaseten United States Zero retention ZDR 20 policy
Boundlessboundless United States Standard no verified privacy claim 20 policy
Google AI Studiogoogle-ai-studio United States Standard no verified privacy claim 16 policy
Anthropicanthropic United States Standard no verified privacy claim 15 policy
Google Vertex AIgoogle-vertex United States Standard ZDR on prepaid 14 policy
Telnyxtelnyx United States Zero retention ZDR 14 policy
Fireworks AIfireworks United States Standard no verified privacy claim 13 policy
Tinfoiltinfoil United States Trusted provider E2EE Confidential ZDR 13 policy
Microsoft Azure AI Foundryazure United States Standard ZDR on prepaid 12 policy
Chuteschutes United States Confidential + E2EE + ZDR provider E2EE Confidential ZDR 11 policy
NVIDIA NIMnvidia-nim United States Standard no verified privacy claim 10 policy
Databricksdatabricks United States Standard no verified privacy claim 9 policy
Waferwafer United States Standard no verified privacy claim 9 policy
xAI Grokgrok United States Standard ZDR on prepaid 9 policy
Crusoecrusoe United States Standard no verified privacy claim 8 policy
DigitalOcean Gradient AIdigitalocean United States Standard no verified privacy claim 7 policy
Togethertogether United States Zero retention ZDR 7 policy
Makoramakora United States Standard no verified privacy claim 6 policy
FriendliAIfriendli United States Standard no verified privacy claim 5 policy
Morphmorph United States Standard no verified privacy claim 5 policy
0G Private Computerzero-g United States No-store Provider TEE claim 4 policy
Decartdecart United States Standard no verified privacy claim 4 policy
ElevenLabselevenlabs United States Standard no verified privacy claim 4 policy
Neurometric AIneurometric United States No-store no verified privacy claim 4 policy
ScaleDownscaledown United States Zero retention ZDR 4 policy
Cerebrascerebras United States Zero retention ZDR 3 policy
CoralBrickscoralbricks United States Standard ZDR on prepaid 3 policy
Metameta United States Standard no verified privacy claim 3 policy
Thinking Machines Labthinkingmachines United States Standard no verified privacy claim 3 policy
Munamuna United States Standard no verified privacy claim 2 policy
Confidential AIconfidential-ai United States Zero retention Provider TEE claim ZDR 1 policy
NEAR AInear-ai United States Confidential + E2EE + ZDR provider E2EE Confidential ZDR 1 policy
Runwayrunway United States Standard no verified privacy claim 1 policy
Voyage AIvoyage United States Zero retention ZDR 1 policy
Lightning AIlightning United States Standard no verified privacy claim 0 policy
TrustedRoutertrustedrouter United States Confidential + E2EE + ZDR provider E2EE Confidential ZDR 0 policy
Vultrvultr United States Standard no verified privacy claim 0 policy

Every provider's full policy note and route list is on its own page. 36 providers in the catalog publish no operator TrustedRouter could pin down, so they appear in no jurisdiction list on this site and no request preference treats them as matching one. Review every provider.

Models built by US labs

193 models

Grouped by the lab that built the weights, with the source TrustedRouter read that from. The operator columns show who runs the endpoints for each route, which is a different question from the lab's country and frequently a different answer.

OpenAI 56 models

OpenAI's terms of use name OpenAI OpCo, LLC, a Delaware company at 1455 3rd Street, San Francisco, CA 94158. Source.

Model Context Routes Operator jurisdictions US-operated routes
openai/gpt-oss-120b
OpenAI: gpt-oss-120b
131,072 38 +10 more on the model page
openai/gpt-oss-20b
OpenAI: gpt-oss-20b
131,072 16 +1 more on the model page
openai/gpt-5.6-sol
OpenAI: GPT-5.6 Sol
1,050,000 5
openai/gpt-5-mini
OpenAI: GPT-5 Mini
400,000 4
openai/gpt-5.5
OpenAI: GPT-5.5
1,050,000 4
openai/gpt-5.6-luna
OpenAI: GPT-5.6 Luna
1,050,000 4

50 further OpenAI models are in the catalog, ranked below these by route count. Browse every model.

Google DeepMind 37 models

Gemini and Gemma come from Google DeepMind, part of Alphabet Inc., whose 10-K gives principal executive offices at 1600 Amphitheatre Parkway, Mountain View, California. The code records that parent's home, not the location of every team that worked on a model. Source.

Model Context Routes Operator jurisdictions US-operated routes
google/gemma-4-31b-it
Google: Gemma 4 31B
262,144 24 +5 more on the model page
google/gemma-4-26b-a4b-it
Google: Gemma 4 26B A4B
262,144 16 +2 more on the model page
google/gemma-3-27b-it
Google: Gemma 3 27B
131,072 12
google/gemma-3-12b-it
Google: Gemma 3 12B
131,072 9
google/gemini-2.5-flash
Google: Gemini 2.5 Flash
1,048,576 7
google/gemini-3.1-flash-lite
Google: Gemini 3.1 Flash Lite
1,048,576 7

31 further Google DeepMind models are in the catalog, ranked below these by route count. Browse every model.

Anthropic 21 models

Anthropic's commercial terms contract through Anthropic, PBC under California law outside the EEA, Switzerland, and the UK, where Anthropic Ireland, Limited applies instead. Source.

Model Context Routes Operator jurisdictions US-operated routes
anthropic/claude-haiku-5.5
Anthropic: Claude Haiku 5.5
1,000,000 4
anthropic/claude-opus-4.8
Anthropic: Claude Opus 4.8
1,000,000 4
anthropic/claude-opus-5.5
Anthropic: Claude Opus 5.5
1,000,000 4
anthropic/claude-sonnet-4.6
Anthropic: Claude Sonnet 4.6
1,000,000 4
anthropic/claude-sonnet-5
Anthropic: Claude Sonnet 5
1,000,000 4
anthropic/claude-sonnet-5.5
Anthropic: Claude Sonnet 5.5
1,000,000 4

15 further Anthropic models are in the catalog, ranked below these by route count. Browse every model.

Meta 21 models

The Llama licence names Meta Platforms, Inc. for licensees outside the EEA and Switzerland, and Meta Platforms Ireland Limited within them. Source.

Model Context Routes Operator jurisdictions US-operated routes
meta-llama/llama-3.3-70b-instruct
Meta: Llama 3.3 70B Instruct
128,000 13 +1 more on the model page
meta-llama/llama-4-scout-17b-16e-instruct
Llama 4 Scout Instruct
131,072 5
meta-llama/llama-3.1-8b-instruct
Meta: Llama 3.1 8B Instruct
131,072 4
meta-models/muse-spark-open
Muse Spark Open
1,048,576 4
meta-llama/llama-4-maverick
Meta: Llama 4 Maverick
524,288 3
meta-llama/llama-3.2-3b
Llama 3.2 3B
128,000 2

15 further Meta models are in the catalog, ranked below these by route count. Browse every model.

NVIDIA 14 models

The Nemotron family comes from NVIDIA Corporation, whose corporate address for stockholder communications is 2788 San Tomas Expressway, Santa Clara, California 95051. Source.

Model Context Routes Operator jurisdictions US-operated routes
nvidia/nemotron-3-ultra-550b-a55b
NVIDIA: Nemotron 3 Ultra
262,144 9
nvidia/nemotron-3.5-super-vl
NVIDIA: Nemotron 3.5 Super VL
1,048,576 7
nvidia/nemotron-3-super-120b-a12b
NVIDIA: Nemotron 3 Super
262,144 5
nvidia/nemotron-3.5-lightning
NVIDIA: Nemotron 3.5 Lightning
262,144 5
nvidia/nemotron-3-nano-30b-a3b
NVIDIA: Nemotron 3 Nano 30B A3B
262,144 4
nvidia/nemotron-3.5-lightning-30b-a3b
nvidia/Nemotron-3.5-Lightning-30B-A3B
262,144 4

8 further NVIDIA models are in the catalog, ranked below these by route count. Browse every model.

xAI 13 models

xAI's consumer terms take legal notices at 1450 Page Mill Rd., Palo Alto, CA 94304 and are governed by the laws of the State of Texas. Source.

Model Context Routes Operator jurisdictions US-operated routes
x-ai/grok-4.20-reasoning
grok-4.20-0309-reasoning
0 4
x-ai/grok-4.3
SpaceXAI: Grok 4.3
1,000,000 3
x-ai/grok-4.5
SpaceXAI: Grok 4.5
500,000 3
x-ai/grok-4.6
SpaceXAI: Grok 4.6
500,000 3
x-ai/grok-build-0.1
SpaceXAI: Grok Build 0.1
256,000 3
x-ai/grok-4.1-fast-non-reasoning
grok-4-1-fast-non-reasoning
0 2

7 further xAI models are in the catalog, ranked below these by route count. Browse every model.

Recraft 8 models

Recraft's terms name Recraft Inc. and give its notices address as 450 Townsend Street, San Francisco, California 94107. Source.

Model Context Routes Operator jurisdictions US-operated routes
recraft/recraftv2
Recraft V2
0 1 none configured
recraft/recraftv3
Recraft V3
0 1 none configured
recraft/recraftv4
Recraft V4
0 1 none configured
recraft/recraftv4_1
Recraft V4.1
0 1 none configured
recraft/recraftv4_1_pro
Recraft V4.1 Pro
0 1 none configured
recraft/recraftv4_1_utility
Recraft V4.1 Utility
0 1 none configured

2 further Recraft models are in the catalog, ranked below these by route count. Browse every model.

Microsoft Research 5 models

The Phi family is built by Microsoft Research. Microsoft's privacy statement identifies Microsoft Corporation at One Microsoft Way, Redmond, Washington 98052, United States. Source.

Model Context Routes Operator jurisdictions US-operated routes
microsoft/phi-4
Microsoft: Phi 4
16,384 2
microsoft/wizardlm-2-8x22b
WizardLM 2 8x22B
65,535 2
microsoft/mai-voice-2.1
Microsoft: MAI-Voice-2.1
0 1
microsoft/mai-voice-2.1-flash
Microsoft: MAI-Voice-2.1-Flash
0 1
microsoft/microsoft-decision-1
Microsoft Decision-1
32,768 1

ElevenLabs 4 models

Eleven Labs Inc. lists its legal address in New York; this is not an inference-residency claim. Source.

Model Context Routes Operator jurisdictions US-operated routes
elevenlabs/eleven-flash-v2.5
ElevenLabs: Eleven Flash V2.5
0 1
elevenlabs/eleven-multilingual-v2
ElevenLabs: Eleven Multilingual V2
0 1
elevenlabs/eleven-turbo-v2.5
ElevenLabs: Eleven Turbo V2.5
0 1
elevenlabs/eleven-v3
ElevenLabs: Eleven V3
0 1

ScaleDown 4 models

Carbon ScaleDown, Inc., a Delaware corporation, publishes these task-specific models. This identifies the publisher, not a claim that its undisclosed underlying model weights are open or US-origin. Source.

Model Context Routes Operator jurisdictions US-operated routes
scaledown/classify
ScaleDown Classify
1,000,000 1
scaledown/compress
ScaleDown Compress
1,000,000 1
scaledown/extract
ScaleDown Extract
1,000,000 1
scaledown/summarize
ScaleDown Summarize
1,000,000 1

Reflection AI 3 models

Reflection AI, Inc.'s August 28, 2026 privacy policy lists its contact address in New York. Its first-party Beam announcement (https://reflection.ai/blog/introducing-beam) identifies the lab as the model's developer. This describes the developer, not inference residency. Source.

Model Context Routes Operator jurisdictions US-operated routes
reflection/beam
Reflection Beam 1.0
1,048,576 2 none configured
reflection/beam-1.0
Beam
1,048,576 2
reflection/beam-1.0-501b-a23b
Reflection: Beam 1.0 501B A23B
1,048,576 2

Thinking Machines Lab 3 models

Thinking Machines Lab Inc. is a public benefit corporation based in San Francisco, California. Its own site names the company without stating a location. Source.

Model Context Routes Operator jurisdictions US-operated routes
thinkingmachines/inkling-2-small
Thinking Machines: Inkling 2 Small
1,048,576 5
thinkingmachines/inkling-small
Thinking Machines: Inkling Small
524,288 4
thinkingmachines/inkling
Thinking Machines: Inkling
524,288 2

IBM Research 2 models

IBM Research identifies Granite 4.2 as its model family. IBM's corporate contact page lists International Business Machines Corporation in Armonk, New York, United States (https://www.ibm.com/contact/global). Source.

Model Context Routes Operator jurisdictions US-operated routes
ibm-granite/granite-4.2-8b
IBM: Granite 4.2 8B
131,072 3
ibm-granite/granite-4.1-8b
granite-4.1-8b
32,768 1

Inception Labs 1 model

Inception AI, Inc.'s terms list its address in Redwood City, California. This describes the model developer, not inference residency. Source.

Model Context Routes Operator jurisdictions US-operated routes
inception/mercury-3
Inception: Mercury 3
1,000,000 1 none configured

TrustedRouter 1 model

First-party record: the Liberty, Athena, OpenPatcher, and orchestration families are built by TrustedRouter, a US company. The OpenPatcher, Athena and Liberty 4.x ids in US_PROVIDER_ONLY_MODEL_IDS additionally force US-operated provider routes at request time. Source.

Model Context Routes Operator jurisdictions US-operated routes
trustedrouter/fortanix-test
Fortanix internal test
131,072 1 none configured

Model origin is keyed by the vendor prefix of a model id, and a prefix earns a row once at least three of its models are in the catalog. 44 callable catalog models sit under a prefix below that line and carry no recorded origin, so they are on no page here rather than assigned a country by guesswork.

Browse every modelReview provider policiesAI models from the European UnionAI models from China

Origin

A lab's country is fixed.

Where a model was built does not change when you route it differently. It is a procurement fact, and it is the one this page's lab sections answer, each with the primary source it was read from.

Traffic

Your route is yours to pick.

Where prompts go is decided per request. provider.jurisdiction accepts us today; provider.only pins an exact allowlist for any other requirement, and both fail closed rather than falling back.

Scope

What this page will not say.

Not that a US operator keeps data in the US, not that a US lab's model is safer, and not that any of it is certified. It reports two recorded fields and the source behind each. Residency is a separate control.

Put these guides into practice

Choose a route, check its privacy requirements, and test it with your own workload.

Sign up today and start with up to 1 million free tokens. No card required. Eligible new email or social signups receive starter credit for their first workspace; token volume depends on model pricing and input/output usage. Wallet-only signups and additional workspaces do not receive this grant.

Live catalog evidence

Current routes, prices, privacy, and measured performance.

Catalog facts come from the routes currently configured in TrustedRouter. Performance uses the same cached metadata snapshot as the public leaderboard. Prompts and outputs are not part of these measurements.

599public models
116providers
1849configured routes
466ZDR routes
53provider E2EE routes
5867recent availability samples
Model Providers Context Input Output Privacy Measured route
Google: Gemini 3.8 Flashgoogle/gemini-3.8-flash
6 routes
1,048,576 $0.79125/1M $3.95625/1M varies 2109 ms TTFT google-vertex · 100.00% available · n=87
Meta: Llama 3.3 70B Instructmeta-llama/llama-3.3-70b-instruct
+5
13 routes
128,000 $0.142425/1M to $1.84625/1M $0.422/1M to $2.90125/1M Confidential + ZDR 6 cited scores 944 ms TTFT akashml · 100.00% available · n=21
Google: Gemini 4 Carbongoogle/gemini-4-carbon
4 routes
1,048,576 $2.11/1M $10.55/1M varies 4693 ms TTFT google-vertex · 42.86% available · n=11
Nemotron 3.5 Super VLnvidia/nemotron-3.5-super-vl-120b-a12b
4 routes
1,048,576 $0.15825/1M to $2.11/1M $0.5275/1M to $10.55/1M ZDR 956 ms TTFT nvidia-nim · 100.00% available · n=11
nvidia/Nemotron-3.5-Lightning-30B-A3Bnvidia/nemotron-3.5-lightning-30b-a3b
4 routes
262,144 $0.05275/1M to $2.11/1M $0.211/1M to $10.55/1M varies 772 ms TTFT nvidia-nim · 100.00% available · n=8
Nano Banana 2google/gemini-3.1-flash-image-preview
2 routes
65,536 $0.5275/1M $63.3/1M varies 1349 ms TTFT google-ai-studio · 100.00% available · n=7

Provider summaries cover each provider's full catalog.

Google AI StudioPolicy varies 16 models 2055 ms p50 · n=71
Google Vertex AIZDR on prepaid 14 models 2109 ms p50 · n=75
GMI CloudPolicy varies 68 models 2608 ms p50 · n=25
Atlas CloudPolicy varies 69 models 2071 ms p50 · n=22
ParasailZDR 25 models 983 ms p50 · n=44
TinfoilConfidential + ZDR 13 models 893 ms p50 · n=497
Novita AIPolicy varies 90 models 1789 ms p50 · n=23
TelnyxZDR 14 models 941 ms p50 · n=81

Browse every modelReview provider policiesOpen the full leaderboardSnapshot 2026-11-19T14:13:58.508Z

Questions

What counts as a US AI model?

Two different things, so TrustedRouter records them separately. A model's origin is the country of the lab that built the weights, read from that lab's own licence, terms, or regulatory filing. A route's jurisdiction is the country of the company operating the endpoint the request reaches, read from that company's own terms or filing. A model from a US lab is regularly served by providers registered outside the US, and open-weights models from other countries are regularly served by US providers, so neither fact implies the other.

How do I make sure my prompts only reach US providers?

Set provider.jurisdiction to "us" on the request. The router then considers only providers whose recorded operator country is the United States and fails closed when none qualify, rather than falling back to a provider you did not approve. For an exact list rather than a country test, use provider.only with the provider slugs you approved.

Does a US provider mean my data stays in the United States?

No, and this site does not claim it. The recorded country is the legal home of the company operating the endpoint, taken from its published terms, privacy policy, or regulatory filing. It says nothing about which datacentre answers a given request. Where processing location is a contractual requirement, it comes from an agreed provider allowlist and a signed agreement, not from a country code in a catalog.

What is the Liberty model family?

Liberty is TrustedRouter's own set of panel routes: each id calls several component models and returns one answer. Every component model under every Liberty route resolves to a lab recorded as US-based in the catalog's model-origin table, and the US-AI-models page shows that as a count computed when the page renders rather than as a fixed claim. US-operated provider routes are available for each Liberty id; the default candidate pool also includes operators outside the US, so add provider.jurisdiction="us" when US-operated serving is a requirement.

Workspace access

Sign in

Choose a sign in method to access your TrustedRouter workspace.

New to TrustedRouter?

By continuing you agree to the terms of service and privacy policy.