Supported GitLab Duo Self-Hosted models and hardware requirements

  • Tier: Premium, Ultimate
  • Add-on: GitLab Duo Enterprise
  • Offering: GitLab Self-Managed

GitLab Duo Self-Hosted supports integration with industry-leading models from Mistral, Meta, Anthropic, and OpenAI through your preferred serving platform.

You can choose from these supported models to match your specific performance needs and use cases.

In GitLab 18.3 and later, you can also bring your own compatible model, giving you the flexibility to experiment with additional language models beyond the officially supported options.

Supported models

Support for the following GitLab-supported large language models (LLMs) is generally available.

  • Fully compatible: The model can likely handle the feature without any loss of quality.
  • Largely compatible: The model supports the feature, but there might be compromises or limitations.
  • Not compatible: The model is unsuitable for the feature, likely resulting in significant quality loss or performance issues. Models that are marked not compatible for a feature will not receive GitLab support for that specific feature.
Model familyModelSupported platformsCode completionCode generationGitLab Duo Chat
Mistral CodestralCodestral 22B v0.1vLLMcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatibleNot applicable
MistralMistral 7B-it v0.3 1vLLMcheck-circle-dashed Largely compatiblecheck-circle-filled Fully compatibledash-circle Not compatible
MistralMixtral 8x7B-it v0.1 1vLLM, AWS Bedrockcheck-circle-dashed Largely compatiblecheck-circle-filled Fully compatiblecheck-circle-dashed Largely compatible
MistralMixtral 8x22B-it v0.1 1vLLMcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatiblecheck-circle-dashed Largely compatible
MistralMistral Small 24B Instruct 2506vLLMcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatiblecheck-circle-filled Fully compatible
Claude 3Claude 3.5 SonnetAWS Bedrockcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatiblecheck-circle-filled Fully compatible
Claude 3Claude 3.7 SonnetAWS Bedrockcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatiblecheck-circle-filled Fully compatible
Claude 4Claude 4 SonnetAWS Bedrockcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatiblecheck-circle-filled Fully compatible
GPTGPT-4 TurboAzure OpenAIcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatiblecheck-circle-dashed Largely compatible
GPTGPT-4oAzure OpenAIcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatiblecheck-circle-filled Fully compatible
GPTGPT-4o-miniAzure OpenAIcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatiblecheck-circle-dashed Largely compatible
LlamaLlama 3 8BvLLMcheck-circle-dashed Largely compatiblecheck-circle-filled Fully compatibledash-circle Not compatible
LlamaLlama 3.1 8BvLLMcheck-circle-dashed Largely compatiblecheck-circle-filled Fully compatiblecheck-circle-dashed Largely compatible
LlamaLlama 3 70BvLLMcheck-circle-dashed Largely compatiblecheck-circle-filled Fully compatibledash-circle Not compatible
LlamaLlama 3.1 70BvLLMcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatiblecheck-circle-filled Fully compatible
LlamaLlama 3.3 70BvLLMcheck-circle-filled Fully compatiblecheck-circle-filled Fully compatiblecheck-circle-filled Fully compatible

Footnotes:

  1. This model is scheduled for deprecation in GitLab 18.5. Mistral Small 24B Instruct 2506 is the recommended alternative.

Bring your own compatible model

  • Status: Beta

You can bring your own compatible models to use with GitLab Duo features.

The general model family provides support for compatible models and platforms that adhere to the OpenAI API specification. Use this model family to try language models that are not explicitly supported by GitLab.

This feature is in beta and is therefore subject to change as we gather feedback and improve the integration:

  • GitLab does not provide technical support for issues specific to your chosen model or platform.
  • Not all GitLab Duo features are guaranteed to work optimally with every compatible model.
  • Response quality, speed, and performance overall might vary significantly based on your model choice.
Model FamilyModel RequirementsSupported PlatformsCode completionCode generationGitLab Duo Chat
GeneralAny model compatible with the OpenAI API specificationAny platform that provides OpenAI-compatible API endpointscheck-circle-dashed Betacheck-circle-dashed Betacheck-circle-dashed Beta

Experimental and beta models

The following models are configurable for the functionalities marked below, but are currently in beta or experimental status, under evaluation, and are excluded from the “Customer Integrated Models” definition in the AI Functionality Terms:

Model familyModelSupported platformsStatusCode completionCode generationGitLab Duo Chat
OpenAI GPTGPT OSS 20bvLLM, AWS Bedrock, Azure OpenAIExperimentalcheck-circle Yesdotted-circle Yesdotted-circle Yes
OpenAI GPTGPT OSS 120bvLLM, AWS Bedrock, Azure OpenAIExperimentalcheck-circle Yesdotted-circle Yesdotted-circle Yes
CodeGemmaCodeGemma 2bvLLMExperimentalcheck-circle Yesdotted-circle Nodotted-circle No
CodeGemmaCodeGemma 7b-itvLLMExperimentaldotted-circle Nocheck-circle Yesdotted-circle No
CodeGemmaCodeGemma 7b-codevLLMExperimentalcheck-circle Yesdotted-circle Nodotted-circle No
Code LlamaCode-Llama 13bvLLMExperimentaldotted-circle Nocheck-circle Yesdotted-circle No
DeepSeek CoderDeepSeek Coder 33b InstructvLLMExperimentalcheck-circle Yescheck-circle Yesdotted-circle No
DeepSeek CoderDeepSeek Coder 33b BasevLLMExperimentalcheck-circle Yesdotted-circle Nodotted-circle No
MistralMistral 7B-it v0.2vLLM
AWS Bedrock
Experimentalcheck-circle Yescheck-circle Yescheck-circle Yes

GitLab AI vendor models

  • Status: Beta

The availability of this feature is controlled by a feature flag. For more information, see the history.

GitLab AI vendor models integrate with GitLab-hosted AI gateway infrastructure to provide access to AI models curated and made available by GitLab. Instead of using your own self-hosted models, you can choose to use GitLab AI vendor models for specific GitLab Duo features.

To choose which features use GitLab AI vendor models, see Configure GitLab AI vendor models.

When enabled for a specific feature:

  • All calls to those features configured with a GitLab AI vendor model use the GitLab-hosted AI gateway, not the self-hosted AI gateway.
  • No detailed logs are generated in the GitLab-hosted AI gateway, even when AI logs are enabled. This prevents unintended leaks of sensitive information.

Hardware requirements

The following hardware specifications are the minimum requirements for running GitLab Duo Self-Hosted on-premise. Requirements vary significantly based on the model size and intended usage:

Base system requirements

  • CPU:
    • Minimum: 8 cores (16 threads)
    • Recommended: 16+ cores for production environments
  • RAM:
    • Minimum: 32 GB
    • Recommended: 64 GB for most models
  • Storage:
    • SSD with sufficient space for model weights and data.

GPU requirements by model size

Model sizeMinimum GPU configurationMinimum VRAM required
7B models
(for example, Mistral 7B)
1x NVIDIA A100 (40 GB)35 GB
22B models
(for example, Codestral 22B)
2x NVIDIA A100 (80 GB)110 GB
Mixtral 8x7B2x NVIDIA A100 (80 GB)220 GB
Mixtral 8x22B8x NVIDIA A100 (80 GB)526 GB

Use Hugging Face’s memory utility to verify memory requirements.

Response time by model size and GPU

Small machine

With a a2-highgpu-2g (2x Nvidia A100 40 GB - 150 GB vRAM) or equivalent:

Model nameNumber of requestsAverage time per request (sec)Average tokens in responseAverage tokens per second per requestTotal time for requestsTotal TPS
Mistral-7B-Instruct-v0.317.09717.0101.197.09101.17
Mistral-7B-Instruct-v0.3108.41764.290.3513.70557.80
Mistral-7B-Instruct-v0.310013.97693.2349.1720.813331.59

Medium machine

With a a2-ultragpu-4g (4x Nvidia A100 40 GB - 340 GB vRAM) machine on GCP or equivalent:

Model nameNumber of requestsAverage time per request (sec)Average tokens in responseAverage tokens per second per requestTotal time for requestsTotal TPS
Mistral-7B-Instruct-v0.313.80499.0131.253.80131.23
Mistral-7B-Instruct-v0.3106.00740.6122.858.19904.22
Mistral-7B-Instruct-v0.310011.71695.7159.0615.544477.34
Mixtral-8x7B-Instruct-v0.116.50400.061.556.5061.53
Mixtral-8x7B-Instruct-v0.11016.58768.940.3332.56236.13
Mixtral-8x7B-Instruct-v0.110025.90767.3826.8755.571380.68

Large machine

With a a2-ultragpu-8g (8 x NVIDIA A100 80 GB - 1360 GB vRAM) machine on GCP or equivalent:

Model nameNumber of requestsAverage time per request (sec)Average tokens in responseAverage tokens per second per requestTotal time for requests (sec)Total TPS
Mistral-7B-Instruct-v0.313.23479.0148.413.22148.36
Mistral-7B-Instruct-v0.3104.95678.3135.986.85989.11
Mistral-7B-Instruct-v0.310010.14713.2769.6313.965108.75
Mixtral-8x7B-Instruct-v0.116.08709.0116.696.07116.64
Mixtral-8x7B-Instruct-v0.1109.95645.063.6813.40481.06
Mixtral-8x7B-Instruct-v0.110013.83585.0141.8020.382869.12
Mixtral-8x22B-Instruct-v0.1114.39828.057.5614.3857.55
Mixtral-8x22B-Instruct-v0.11020.57629.730.2428.02224.71
Mixtral-8x22B-Instruct-v0.110027.58592.4921.3436.801609.85

AI Gateway Hardware Requirements

For recommendations on AI gateway hardware, see the AI gateway scaling recommendations.