Red Hat Summit Red Hat SummitSupportConsoleDevelopersStart a trialContact

Validated models

Red Hat AI 3

Red Hat AI validated models

Red Hat AI Documentation Team

Abstract

Learn about the validated models that you can inference serve with Red Hat AI.

Preface
Copy link

Red Hat provides validated models that you can inference serve with Red Hat AI.

Chapter 1. Red Hat AI validated models
Copy link

The Red Hat AI models are validated using open source tools. The model type you use depends on how you want to deploy the model.

Note

If you are using AI Inference Server as part of a RHEL AI deployment, use OCI artifact images.

If you are using AI Inference Server as part of a OpenShift AI deployment, use ModelCar images.

Red Hat uses GuideLLM for performance benchmarking and Language Model Evaluation Harness for accuracy evaluations.

Explore the Red Hat AI validated models collections on Hugging Face.

Important

AMD GPUs support FP8 (W8A8) and GGUF quantization variant models only. For more information, see Supported hardware.

The following models, available from RedHat AI on Hugging Face, are validated for use with Red Hat AI Inference Server.

Expand
Table 1.1. Red Hat AI validated models - October 2025 collection
ModelQuantized variantsHugging Face model cardValidated on

gpt-oss-120b

None

  • RHAIIS 3.2.2
  • RHOAI 2.25

gpt-oss-20b

None

  • RHAIIS 3.2.2
  • RHOAI 2.25

NVIDIA-Nemotron-Nano-9B-v2

INT4, FP8

  • RHAIIS 3.2.2
  • RHOAI 2.25

Qwen3-Coder-480B-A35B-Instruct

FP8

  • RHAIIS 3.2.2
  • RHOAI 2.25

Voxtral-Mini-3B-2507

FP8

  • RHAIIS 3.2.2
  • RHOAI 2.25

whisper-large-v3-turbo

INT4

  • RHAIIS 3.2.2
  • RHOAI 2.25

The following models, available from RedHat AI on Hugging Face, are validated for use with Red Hat AI Inference Server.

Expand
Table 1.2. Red Hat AI validated models - September 2025 collection
ModelQuantized variantsHugging Face model cardValidated on

DeepSeek-R1-0528

INT4

  • RHAIIS 3.2.1
  • RHOAI 2.24

gemma-3n-E4B-it

FP8

  • RHAIIS 3.2.1
  • RHOAI 2.24

Kimi-K2-Instruct

INT4

  • RHAIIS 3.2.1
  • RHOAI 2.24

Qwen3-8B

FP8

  • RHAIIS 3.2.1
  • RHOAI 2.24

The following models, available from RedHat AI on Hugging Face, are validated for use with Red Hat AI Inference Server.

Expand
Table 1.3. Red Hat AI validated models - May 2025 collection
ModelQuantized variantsHugging Face model cardValidated on

gemma-2-9b-it

FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

granite-3.1-8b-base

INT4

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

granite-3.1-8b-instruct

INT4, INT8, FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

Llama-3.1-8B-Instruct

None

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

Llama-3.1-Nemotron-70B-Instruct-HF

FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

Llama-3.3-70B-Instruct

INT4, INT8, FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

Llama-4-Maverick-17B-128E-Instruct

FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

Llama-4-Scout-17B-16E-Instruct

INT4, FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

Meta-Llama-3.1-8B-Instruct

INT4, INT8, FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

Mistral-Small-24B-Instruct-2501

INT4, INT8, FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

Mistral-Small-3.1-24B-Instruct-2503

INT4, INT8, FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

Mixtral-8x7B-Instruct-v0.1

None

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

phi-4

INT4, INT8, FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

Qwen2.5-7B-Instruct

INT4, INT8, FP8

  • RHAIIS 3.0
  • RHELAI 1.5
  • RHOAI 2.20

1.4. Validated OCI artifact model container images
Copy link

Expand
Table 1.4. Validated OCI artifact model container images
ModelQuantized variantsModelCar images

llama-4-scout-17b-16e-instruct

INT4, FP8

  • Baseline: registry.redhat.io/rhelai1/llama-4-scout-17b-16e-instruct:1.5
  • INT4: registry.redhat.io/rhelai1/llama-4-scout-17b-16e-instruct-quantized-w4a16:1.5
  • FP8: registry.redhat.io/rhelai1/llama-4-scout-17b-16e-instruct-fp8-dynamic:1.5

llama-4-maverick-17b-128e-instruct

FP8

  • Baseline: registry.redhat.io/rhelai1/llama-4-maverick-17b-128e-instruct:1.5
  • FP8: registry.redhat.io/rhelai1/llama-4-maverick-17b-128e-instruct-fp8:1.5

mistral-small-3-1-24b-instruct-2503

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/mistral-small-3-1-24b-instruct-2503:1.5
  • INT4: registry.redhat.io/rhelai1/mistral-small-3-1-24b-instruct-2503-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/mistral-small-3-1-24b-instruct-2503-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/mistral-small-3-1-24b-instruct-2503-fp8-dynamic:1.5

llama-3-3-70b-instruct

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/llama-3-3-70b-instruct:1.5
  • INT4: registry.redhat.io/rhelai1/llama-3-3-70b-instruct-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/llama-3-3-70b-instruct-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/llama-3-3-70b-instruct-fp8-dynamic:1.5

llama-3-1-8b-instruct

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/llama-3-1-8b-instruct:1.5
  • INT4: registry.redhat.io/rhelai1/llama-3-1-8b-instruct-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/llama-3-1-8b-instruct-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/llama-3-1-8b-instruct-fp8-dynamic:1.5

granite-3-1-8b-instruct

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/granite-3-1-8b-instruct:1.5
  • INT4: registry.redhat.io/rhelai1/granite-3-1-8b-instruct-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/granite-3-1-8b-instruct-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/granite-3-1-8b-instruct-fp8-dynamic:1.5

phi-4

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/phi-4:1.5
  • INT4: registry.redhat.io/rhelai1/phi-4-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/phi-4-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/phi-4-fp8-dynamic:1.5

qwen2-5-7b-instruct

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/qwen2-5-7b-instruct:1.5
  • INT4: registry.redhat.io/rhelai1/qwen2-5-7b-instruct-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/qwen2-5-7b-instruct-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/qwen2-5-7b-instruct-fp8-dynamic:1.5

mistral-small-24b-instruct-2501

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/mistral-small-24b-instruct-2501:1.5
  • INT4: registry.redhat.io/rhelai1/mistral-small-24b-instruct-2501-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/mistral-small-24b-instruct-2501-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/mistral-small-24b-instruct-2501-fp8-dynamic:1.5

mixtral-8x7b-instruct-v0-1

None

  • Baseline: registry.redhat.io/rhelai1/mixtral-8x7b-instruct-v0-1:1.4

granite-3-1-8b-base

INT4 (baseline currently unavailable)

  • INT4: registry.redhat.io/rhelai1/granite-3-1-8b-base-quantized-w4a16:1.5

granite-3.1-8b-starter-v2

None

  • Baseline: registry.redhat.io/rhelai1/granite-3.1-8b-starter-v2:1.5

llama-3-1-nemotron-70b-instruct-hf

FP8

  • Baseline: registry.redhat.io/rhelai1/llama-3-1-nemotron-70b-instruct-hf:1.5
  • FP8: registry.redhat.io/rhelai1/llama-3-1-nemotron-70b-instruct-hf-fp8-dynamic:1.5

gemma-2-9b-it

FP8

  • Baseline: registry.redhat.io/rhelai1/gemma-2-9b-it:1.5
  • FP8: registry.redhat.io/rhelai1/gemma-2-9b-it-fp8:1.5

deepseek-r1-0528

INT4 (baseline currently unavailable)

  • INT4: registry.redhat.io/rhelai1/deepseek-r1-0528-quantized-w4a16:1.5

qwen3-8b

FP8 (baseline currently unavailable)

  • FP8: registry.redhat.io/rhelai1/qwen3-8b-fp8-dynamic:1.5

kimi-k2-instruct

INT4 (baseline currently unavailable)

  • INT4: registry.redhat.io/rhelai1/kimi-k2-instruct-quantized-w4a16:1.5

gemma-3n-e4b-it

FP8 (baseline currently unavailable)

  • FP8: registry.redhat.io/rhelai1/gemma-3n-e4b-it-fp8-dynamic:1.5

gpt-oss-120b

None

  • Baseline: registry.redhat.io/rhelai1/gpt-oss-120b:1.5

gpt-oss-20b

None

  • Baseline: registry.redhat.io/rhelai1/gpt-oss-20b:1.5

qwen3-coder-480b-a35b-instruct

FP8 (baseline currently unavailable)

  • FP8: registry.redhat.io/rhelai1/qwen3-coder-480b-a35b-instruct-fp8:1.5

whisper-large-v3-turbo

INT4 (baseline currently unavailable)

  • INT4: registry.redhat.io/rhelai1/whisper-large-v3-turbo-quantized-w4a16:1.5

voxtral-mini-3b-2507

FP8 (baseline currently unavailable)

  • FP8: registry.redhat.io/rhelai1/voxtral-mini-3b-2507-fp8-dynamic:1.5

nvidia-nemotron-nano-9b-v2

FP8 (baseline currently unavailable)

  • FP8: registry.redhat.io/rhelai1/nvidia-nemotron-nano-9b-v2-fp8-dynamic:1.5

1.5. ModelCar container images
Copy link

Expand
Table 1.5. ModelCar container images
ModelQuantized variantsModelCar images

llama-4-scout-17b-16e-instruct

INT4, FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-llama-4-scout-17b-16e-instruct:1.5
  • INT4: registry.redhat.io/rhelai1/modelcar-llama-4-scout-17b-16e-instruct-quantized-w4a16:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-llama-4-scout-17b-16e-instruct-fp8-dynamic:1.5

llama-4-maverick-17b-128e-instruct

FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-llama-4-maverick-17b-128e-instruct:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-llama-4-maverick-17b-128e-instruct-fp8:1.5

mistral-small-3-1-24b-instruct-2503

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-mistral-small-3-1-24b-instruct-2503:1.5
  • INT4: registry.redhat.io/rhelai1/modelcar-mistral-small-3-1-24b-instruct-2503-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/modelcar-mistral-small-3-1-24b-instruct-2503-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-mistral-small-3-1-24b-instruct-2503-fp8-dynamic:1.5

llama-3-3-70b-instruct

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-llama-3-3-70b-instruct:1.5
  • INT4: registry.redhat.io/rhelai1/modelcar-llama-3-3-70b-instruct-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/modelcar-llama-3-3-70b-instruct-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-llama-3-3-70b-instruct-fp8-dynamic:1.5

llama-3-1-8b-instruct

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-llama-3-1-8b-instruct:1.5
  • INT4: registry.redhat.io/rhelai1/modelcar-llama-3-1-8b-instruct-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/modelcar-llama-3-1-8b-instruct-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-llama-3-1-8b-instruct-fp8-dynamic:1.5

granite-3-1-8b-instruct

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-granite-3-1-8b-instruct:1.5
  • INT4: registry.redhat.io/rhelai1/modelcar-granite-3-1-8b-instruct-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/modelcar-granite-3-1-8b-instruct-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-granite-3-1-8b-instruct-fp8-dynamic:1.5

phi-4

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-phi-4:1.5
  • INT4: registry.redhat.io/rhelai1/modelcar-phi-4-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/modelcar-phi-4-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-phi-4-fp8-dynamic:1.5

qwen2-5-7b-instruct

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-qwen2-5-7b-instruct:1.5
  • INT4: registry.redhat.io/rhelai1/modelcar-qwen2-5-7b-instruct-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/modelcar-qwen2-5-7b-instruct-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-qwen2-5-7b-instruct-fp8-dynamic:1.5

mistral-small-24b-instruct-2501

INT4, INT8, FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-mistral-small-24b-instruct-2501:1.5
  • INT4: registry.redhat.io/rhelai1/modelcar-mistral-small-24b-instruct-2501-quantized-w4a16:1.5
  • INT8: registry.redhat.io/rhelai1/modelcar-mistral-small-24b-instruct-2501-quantized-w8a8:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-mistral-small-24b-instruct-2501-fp8-dynamic:1.5

mixtral-8x7b-instruct-v0-1

None

  • Baseline: registry.redhat.io/rhelai1/modelcar-mixtral-8x7b-instruct-v0-1:1.4

granite-3-1-8b-base

INT4 (baseline currently unavailable)

  • INT4: registry.redhat.io/rhelai1/modelcar-granite-3-1-8b-base-quantized-w4a16:1.5

granite-3-1-8b-starter-v2

None

  • Baseline: registry.redhat.io/rhelai1/modelcar-granite-3-1-8b-starter-v2:1.5

llama-3-1-nemotron-70b-instruct-hf

FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-llama-3-1-nemotron-70b-instruct-hf:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-llama-3-1-nemotron-70b-instruct-hf-fp8-dynamic:1.5

gemma-2-9b-it

FP8

  • Baseline: registry.redhat.io/rhelai1/modelcar-gemma-2-9b-it:1.5
  • FP8: registry.redhat.io/rhelai1/modelcar-gemma-2-9b-it-fp8:1.5

deepseek-r1-0528

INT4 (baseline currently unavailable)

  • INT4: registry.redhat.io/rhelai1/modelcar-deepseek-r1-0528-quantized-w4a16:1.5

qwen3-8b

FP8 (baseline currently unavailable)

  • FP8: registry.redhat.io/rhelai1/modelcar-qwen3-8b-fp8-dynamic:1.5

kimi-k2-instruct

INT4 (baseline currently unavailable)

  • INT4: registry.redhat.io/rhelai1/modelcar-kimi-k2-instruct-quantized-w4a16:1.5

gemma-3n-e4b-it

FP8 (baseline currently unavailable)

  • FP8: registry.redhat.io/rhelai1/modelcar-gemma-3n-e4b-it-fp8-dynamic:1.5

gpt-oss-120b

None

  • Baseline: registry.redhat.io/rhelai1/modelcar-gpt-oss-120b:1.5

gpt-oss-20b

None

  • Baseline: registry.redhat.io/rhelai1/modelcar-gpt-oss-20b:1.5

qwen3-coder-480b-a35b-instruct

FP8 (baseline currently unavailable)

  • FP8: registry.redhat.io/rhelai1/modelcar-qwen3-coder-480b-a35b-instruct-fp8:1.5

whisper-large-v3-turbo

INT4 (baseline currently unavailable)

  • INT4: registry.redhat.io/rhelai1/modelcar-whisper-large-v3-turbo-quantized-w4a16:1.5

voxtral-mini-3b-2507

FP8 (baseline currently unavailable)

  • FP8: registry.redhat.io/rhelai1/modelcar-voxtral-mini-3b-2507-fp8-dynamic:1.5

nvidia-nemotron-nano-9b-v2

FP8 (baseline currently unavailable)

  • FP8: registry.redhat.io/rhelai1/modelcar-nvidia-nemotron-nano-9b-v2-fp8-dynamic:1.5

The following large language models are supported for IBM Power systems with IBM Spyre AI accelerators.

Note

IBM Spyre AI accelerator cards support FP16 format model weights only. For compatible models, the Red Hat AI Inference Server inference engine automatically converts weights to FP16 at startup. No additional configuration is needed.

Expand
Table 2.1. IBM Granite models for use with IBM Spyre AI accelerators
ModelHugging Face model card

granite-3.3-8b-instruct

ibm-granite/granite-3.3-8b-instruct

granite-embedding-125m-english

ibm-granite/granite-embedding-125m-english

granite-embedding-278m-multilingual

ibm-granite/granite-embedding-278m-multilingual

Important

Pre-built IBM Granite models run with the specific Python packages that are included in the Red Hat AI Inference Server Spyre container image. The models are tied to fixed configurations for Spyre card count, batch size, and input/output context sizes.

Updating or replacing Python packages in the Red Hat AI Inference Server Spyre container image is not supported.

The following large language models are supported for IBM Z systems with IBM Spyre AI accelerators.

Note

IBM Spyre AI accelerator cards support FP16 format model weights only. For compatible models, the Red Hat AI Inference Server inference engine automatically converts weights to FP16 at startup. No additional configuration is needed.

Expand
Table 3.1. Decoder models for use with IBM Spyre AI accelerators
ModelHugging Face model card

granite-3.3-8b-instruct

ibm-granite/granite-3.3-8b-instruct

Important

Pre-built IBM Granite models run with the specific Python packages that are included in the Red Hat AI Inference Server Spyre container image. The models are tied to fixed configurations for Spyre card count, batch size, and input/output context sizes.

Updating or replacing Python packages in the Red Hat AI Inference Server Spyre container image is not supported.

Legal Notice
Copy link

Copyright © 2025 Red Hat, Inc.
The text of and illustrations in this document are licensed by Red Hat under a Creative Commons Attribution–Share Alike 3.0 Unported license ("CC-BY-SA"). An explanation of CC-BY-SA is available at http://creativecommons.org/licenses/by-sa/3.0/. In accordance with CC-BY-SA, if you distribute this document or an adaptation of it, you must provide the URL for the original version.
Red Hat, as the licensor of this document, waives the right to enforce, and agrees not to assert, Section 4d of CC-BY-SA to the fullest extent permitted by applicable law.
Red Hat, Red Hat Enterprise Linux, the Shadowman logo, the Red Hat logo, JBoss, OpenShift, Fedora, the Infinity logo, and RHCE are trademarks of Red Hat, Inc., registered in the United States and other countries.
Linux® is the registered trademark of Linus Torvalds in the United States and other countries.
Java® is a registered trademark of Oracle and/or its affiliates.
XFS® is a trademark of Silicon Graphics International Corp. or its subsidiaries in the United States and/or other countries.
MySQL® is a registered trademark of MySQL AB in the United States, the European Union and other countries.
Node.js® is an official trademark of Joyent. Red Hat is not formally related to or endorsed by the official Joyent Node.js open source or commercial project.
The OpenStack® Word Mark and OpenStack logo are either registered trademarks/service marks or trademarks/service marks of the OpenStack Foundation, in the United States and other countries and are used with the OpenStack Foundation's permission. We are not affiliated with, endorsed or sponsored by the OpenStack Foundation, or the OpenStack community.
All other trademarks are the property of their respective owners.
Back to top
Red Hat logoGithubredditYoutubeTwitter

Learn

Try, buy, & sell

Communities

About Red Hat Documentation

We help Red Hat users innovate and achieve their goals with our products and services with content they can trust. Explore our recent updates.

Making open source more inclusive

Red Hat is committed to replacing problematic language in our code, documentation, and web properties. For more details, see the Red Hat Blog.

About Red Hat

We deliver hardened solutions that make it easier for enterprises to work across platforms and environments, from the core datacenter to the network edge.

Theme

© 2025 Red Hat