---
title: "Model providers"
description: "Compare LLM providers, from OpenAI, Anthropic, and Google to open-weight labs such as DeepSeek, Qwen, Mistral, and Kimi, by access, licensing, and platforms."
type: "platform-category"
canonical_url: "https://theaiengineering.com/platforms/model-providers"
markdown_url: "https://theaiengineering.com/platforms/model-providers.md"
topic: "LLM models and benchmarks"
author: "The AI Engineering Editors"
publisher: "The AI Engineering"
citation: "The AI Engineering Editors. \"Model providers.\" The AI Engineering. https://theaiengineering.com/platforms/model-providers"
---

# Model providers

A model provider is a company or research lab that trains large language models and makes them available through a hosted API, downloadable open weights, or both. Providers differ in the modalities they support, how they license their weights, where their models run, and which cloud platforms resell them.

Some providers, such as Anthropic, offer models only through their own API and cloud partners. Others, including DeepSeek, Alibaba's Qwen team, Mistral AI, Moonshot AI, and Z.ai, also publish weights that teams can run on their own hardware, under licenses that range from MIT and Apache-2.0 to custom terms with commercial conditions. US labs such as OpenAI, Reflection AI, and Thinking Machines Lab now publish open weights as well.

This list covers labs that build their own language and multimodal foundation models; image-only and audio-only labs are out of scope. To reach several providers through one endpoint, teams usually add an AI gateway, and independent benchmarks are a better guide to model quality than vendor announcements.

## Model providers compared

| Platform | Open source | Deployment | Best for |
| --- | --- | --- | --- |
| [Ai2 (Olmo)](https://allenai.org) | No | Self-hosted | Fully open models for research and auditing |
| [AI21 Labs](https://www.ai21.com) | No | API, Self-hosted | Custom post-trained models and hybrid-architecture open weights |
| [Alibaba (Qwen)](https://qwen.ai) | No | API, Self-hosted | Open weights across a wide range of model sizes |
| [Amazon Nova](https://nova.amazon.com) | No | API | Amazon's own models for teams building on AWS |
| [Anthropic](https://www.anthropic.com) | No | API | Agentic and coding workloads across several clouds |
| [Arcee AI](https://www.arcee.ai) | No | Self-hosted | US-built open weights for customization |
| [Baidu (ERNIE)](https://ernie.baidu.com) | No | API, Self-hosted | Chinese-language models with open and hosted options |
| [ByteDance Seed](https://seed.bytedance.com) | No | API, Self-hosted | Seed models through ByteDance's cloud platforms |
| [Cohere](https://cohere.com) | No | API, Self-hosted | Enterprise retrieval, reranking, and private deployments |
| [DeepSeek](https://www.deepseek.com) | No | API, Self-hosted | MIT-licensed open weights for self-hosting or API use |
| [Google DeepMind (Gemini and Gemma)](https://ai.google.dev) | No | API, Self-hosted | Multimodal and long-context work in the Google ecosystem |
| [IBM Granite](https://www.ibm.com/granite) | No | API, Self-hosted | Apache-licensed enterprise models with vendor support |
| [Inception (Mercury)](https://www.inceptionlabs.ai) | No | API | Low-latency generation with diffusion LLMs |
| [inclusionAI (Ant Group)](https://www.inclusion-ai.org) | No | Self-hosted | Permissively licensed open models from Ant Group |
| [LG AI Research (EXAONE)](https://www.lgresearch.ai/exaone) | No | Self-hosted | Korean-English models from an industrial research lab |
| [Liquid AI](https://www.liquid.ai) | No | API, Self-hosted, Local | On-device and edge models |
| [Meituan LongCat](https://longcat.ai) | No | API, Self-hosted | MIT-licensed open weights from a consumer tech company |
| [Meta (Muse and Llama)](https://dev.meta.ai) | No | API, Self-hosted | Meta models through an API or as open weights |
| [Microsoft (MAI and Phi)](https://microsoft.ai) | No | API, Self-hosted | Small open models and Microsoft's in-house frontier models |
| [MiniMax](https://www.minimax.io) | No | API, Self-hosted | Open-weight agentic LLMs plus speech and video models |
| [Mistral AI](https://mistral.ai) | No | API, Self-hosted | Self-deployable models from a European provider |
| [Moonshot AI (Kimi)](https://www.moonshot.ai) | No | API, Self-hosted | Open-weight mixture-of-experts models for agents and coding |
| [Nous Research](https://nousresearch.com) | No | Self-hosted | Open-weight Hermes models and agent tooling |
| [NVIDIA Nemotron](https://www.nvidia.com/en-us/ai-data-science/foundation-models/nemotron/) | No | API, Self-hosted | Open models with published data, tuned for NVIDIA hardware |
| [OpenAI](https://openai.com) | No | API, Self-hosted | One API for text, voice, and image models |
| [Reflection AI](https://reflection.ai) | No | API, Self-hosted | Large Apache-licensed open weights from a US lab |
| [Reka](https://reka.ai) | No | API, Self-hosted | Multimodal video and audio understanding |
| [Sarvam AI](https://www.sarvam.ai) | No | API, Self-hosted | Indian-language text and voice applications |
| [SpaceXSI (formerly xAI)](https://x.ai) | No | API | Grok models for text, code, voice, and media |
| [StepFun](https://www.stepfun.com) | No | API, Self-hosted | Open-weight Step models and speech APIs |
| [Tencent Hy (Hunyuan)](https://aistudio.tencent.com) | No | API, Self-hosted | Open-weight text, image, video, and 3D models |
| [Thinking Machines Lab](https://thinkingmachines.ai) | No | API, Self-hosted | Fine-tuning open-weight models through a managed API |
| [TII (Falcon)](https://falconllm.tii.ae) | No | Self-hosted | Open-weight models from a government research institute |
| [TypeSafe AI](https://typesafe.ai) | No | API | Typed decisions inside automation workflows |
| [Upstage (Solar)](https://www.upstage.ai) | No | API, Self-hosted | Korean and Japanese models and document parsing |
| [Writer (Palmyra)](https://writer.com) | No | API | Enterprise agents on a vendor's own models |
| [Xiaomi MiMo](https://mimo.xiaomi.com) | No | API, Self-hosted | MIT-licensed reasoning and coding models |
| [Z.ai (GLM)](https://z.ai) | No | API, Self-hosted | MIT-licensed GLM models for coding and agents |

## Profiles

### Ai2 (Olmo)

Ai2, the nonprofit research institute founded by Paul Allen, releases fully open models such as Olmo and Molmo together with their training data, code, and checkpoints, mostly under the Apache-2.0 license.

- Company: Allen Institute for AI
- Website: https://allenai.org
- GitHub: https://github.com/allenai
- Headquarters: Seattle, United States
- Founded: 2014
- Open source: No
- Deployment: Self-hosted
- Best for: Fully open models for research and auditing
- Permalink: https://theaiengineering.com/platforms/model-providers#ai2

### AI21 Labs

AI21 Labs developed the Jamba family of hybrid Transformer-Mamba models, with Jamba2 weights released under Apache-2.0, and now centers its business on post-training specialized models for enterprises with forward-deployed researchers.

- Website: https://www.ai21.com
- GitHub: https://github.com/AI21Labs
- Headquarters: Tel Aviv, Israel
- Founded: 2017
- Open source: No
- Deployment: API, Self-hosted
- Best for: Custom post-trained models and hybrid-architecture open weights
- Permalink: https://theaiengineering.com/platforms/model-providers#ai21-labs

### Alibaba (Qwen)

Alibaba's Qwen team releases open-weight language, coding, vision, and image models, many under the Apache-2.0 license, and offers additional API-only models through Alibaba Cloud Model Studio.

- Company: Alibaba Cloud
- Website: https://qwen.ai
- GitHub: https://github.com/QwenLM
- Headquarters: Hangzhou, China
- Open source: No
- Deployment: API, Self-hosted
- Best for: Open weights across a wide range of model sizes
- Permalink: https://theaiengineering.com/platforms/model-providers#qwen

### Amazon Nova

Amazon Nova is Amazon's family of foundation models, including Nova 2 Lite and Nova 2 Sonic, offered through Amazon Bedrock and a Nova API key, with Nova Forge for building custom models and Nova Act for browser agents.

- Company: Amazon
- Website: https://nova.amazon.com
- Open source: No
- Deployment: API
- Best for: Amazon's own models for teams building on AWS
- Permalink: https://theaiengineering.com/platforms/model-providers#amazon-nova

### Anthropic

Anthropic develops the Claude model family, available through the Claude API and on Amazon Bedrock, Google Cloud, and Microsoft Foundry, with support for tool use, prompt caching, extended thinking, and computer use.

- Website: https://www.anthropic.com
- GitHub: https://github.com/anthropics
- Headquarters: San Francisco, United States
- Founded: 2021
- Open source: No
- Deployment: API
- Languages: Python, TypeScript, Go, Java, Ruby, C#
- Best for: Agentic and coding workloads across several clouds
- Permalink: https://theaiengineering.com/platforms/model-providers#anthropic

### Arcee AI

Arcee AI is a US-based open model lab that builds the Trinity family in Nano, Mini, and Large sizes, released as open weights designed to be transparent, customizable, and deployable on the infrastructure a team chooses.

- Website: https://www.arcee.ai
- GitHub: https://github.com/arcee-ai
- Open source: No
- Deployment: Self-hosted
- Best for: US-built open weights for customization
- Permalink: https://theaiengineering.com/platforms/model-providers#arcee-ai

### Baidu (ERNIE)

Baidu develops the ERNIE model family, offers it through the Qianfan platform on Baidu AI Cloud, and has released open-weight ERNIE models, including the ERNIE 4.5 family under the Apache-2.0 license.

- Company: Baidu
- Website: https://ernie.baidu.com
- GitHub: https://github.com/PaddlePaddle/ERNIE
- Headquarters: Beijing, China
- Open source: No
- Deployment: API, Self-hosted
- Best for: Chinese-language models with open and hosted options
- Permalink: https://theaiengineering.com/platforms/model-providers#baidu-ernie

### ByteDance Seed

ByteDance Seed, the company's AI research team established in 2023, develops the Seed language, vision, speech, and video models, serves them through Volcano Engine and BytePlus ModelArk, and publishes open-weight releases such as Seed-OSS.

- Company: ByteDance
- Website: https://seed.bytedance.com
- GitHub: https://github.com/ByteDance-Seed
- Founded: 2023
- Open source: No
- Deployment: API, Self-hosted
- Best for: Seed models through ByteDance's cloud platforms
- Permalink: https://theaiengineering.com/platforms/model-providers#bytedance-seed

### Cohere

Cohere builds enterprise models, including the Command generative models and the Embed and Rerank retrieval models, and offers them through its API, the North agent platform, Model Vault dedicated inference, and private deployments.

- Website: https://cohere.com
- GitHub: https://github.com/cohere-ai
- Headquarters: Toronto, Canada
- Founded: 2019
- Open source: No
- Deployment: API, Self-hosted
- Languages: Python, TypeScript, Go, Java
- Best for: Enterprise retrieval, reranking, and private deployments
- Permalink: https://theaiengineering.com/platforms/model-providers#cohere

### DeepSeek

DeepSeek, an AI lab based in Hangzhou, publishes its DeepSeek-V4 language models and earlier R1 reasoning models as open weights under the MIT license and serves them through an OpenAI-compatible API.

- Website: https://www.deepseek.com
- GitHub: https://github.com/deepseek-ai
- Headquarters: Hangzhou, China
- Founded: 2023
- Open source: No
- Deployment: API, Self-hosted
- Best for: MIT-licensed open weights for self-hosting or API use
- Permalink: https://theaiengineering.com/platforms/model-providers#deepseek

### Google DeepMind (Gemini and Gemma)

Google DeepMind builds the Gemini models, which Google offers through the Gemini API, Google AI Studio, and Gemini Enterprise Agent Platform on Google Cloud, and it publishes the Gemma family of open-weight models.

- Company: Google
- Website: https://ai.google.dev
- GitHub: https://github.com/google-gemini
- Headquarters: London, United Kingdom
- Open source: No
- Deployment: API, Self-hosted
- Languages: Python, TypeScript, Go, Java
- Best for: Multimodal and long-context work in the Google ecosystem
- Permalink: https://theaiengineering.com/platforms/model-providers#google-gemini

### IBM Granite

IBM Granite is a family of open models for business that covers language, vision, speech, time series, and guardrail tasks, released under the Apache-2.0 license and available on watsonx.ai, Hugging Face, and other platforms.

- Company: IBM
- Website: https://www.ibm.com/granite
- GitHub: https://github.com/ibm-granite
- Headquarters: Armonk, United States
- Open source: No
- Deployment: API, Self-hosted
- Best for: Apache-licensed enterprise models with vendor support
- Permalink: https://theaiengineering.com/platforms/model-providers#ibm-granite

### Inception (Mercury)

Inception develops Mercury, a family of diffusion large language models that generate many tokens in parallel instead of one at a time, offered through its API and enterprise deployments for latency-sensitive applications.

- Company: Inception Labs
- Website: https://www.inceptionlabs.ai
- GitHub: https://github.com/inception-labs
- Open source: No
- Deployment: API
- Best for: Low-latency generation with diffusion LLMs
- Permalink: https://theaiengineering.com/platforms/model-providers#inception-mercury

### inclusionAI (Ant Group)

inclusionAI, the open AI initiative backed by Ant Group, develops the Ling language models and related reasoning and multimodal models, and publishes most of them as open weights under the MIT or Apache-2.0 license.

- Company: Ant Group
- Website: https://www.inclusion-ai.org
- GitHub: https://github.com/inclusionAI
- Headquarters: Hangzhou, China
- Open source: No
- Deployment: Self-hosted
- Best for: Permissively licensed open models from Ant Group
- Permalink: https://theaiengineering.com/platforms/model-providers#inclusionai

### LG AI Research (EXAONE)

LG AI Research develops EXAONE, a family of Korean and English language models, publishing open weights that include the K-EXAONE 2.0 series under Apache-2.0 and earlier EXAONE versions under LG's own model license.

- Company: LG AI Research
- Website: https://www.lgresearch.ai/exaone
- GitHub: https://github.com/LG-AI-EXAONE
- Headquarters: Seoul, South Korea
- Open source: No
- Deployment: Self-hosted
- Best for: Korean-English models from an industrial research lab
- Permalink: https://theaiengineering.com/platforms/model-providers#lg-exaone

### Liquid AI

Liquid AI builds Liquid Foundation Models (LFMs), compact open-weight language, vision, and audio models designed for low latency on phones, laptops, and other edge hardware as well as in the cloud.

- Website: https://www.liquid.ai
- GitHub: https://github.com/Liquid4All
- Headquarters: Cambridge, Massachusetts, United States
- Founded: 2023
- Open source: No
- Deployment: API, Self-hosted, Local
- Best for: On-device and edge models
- Permalink: https://theaiengineering.com/platforms/model-providers#liquid-ai

### Meituan LongCat

Meituan's LongCat team develops LongCat language, reasoning, audio, and video models, releases them as open weights under the MIT license, and offers them through LongCat AI products, including a coding agent.

- Company: Meituan
- Website: https://longcat.ai
- GitHub: https://github.com/meituan-longcat
- Headquarters: Beijing, China
- Open source: No
- Deployment: API, Self-hosted
- Best for: MIT-licensed open weights from a consumer tech company
- Permalink: https://theaiengineering.com/platforms/model-providers#meituan-longcat

### Meta (Muse and Llama)

Meta serves its Muse models, including Muse Spark and Muse Wonder, through the OpenAI-compatible Meta Model API, and publishes open-weight models such as Muse Spark Open and the Llama family under Meta's own licenses.

- Company: Meta Platforms
- Website: https://dev.meta.ai
- GitHub: https://github.com/meta-llama
- Headquarters: Menlo Park, United States
- Open source: No
- Deployment: API, Self-hosted
- Best for: Meta models through an API or as open weights
- Permalink: https://theaiengineering.com/platforms/model-providers#meta

### Microsoft (MAI and Phi)

Microsoft AI develops the MAI models for reasoning, coding, cybersecurity, voice, and transcription, used in Copilot and GitHub Copilot, while Microsoft also publishes the Phi small language models as open weights, many under the MIT license.

- Company: Microsoft
- Website: https://microsoft.ai
- GitHub: https://github.com/microsoft
- Headquarters: Redmond, United States
- Open source: No
- Deployment: API, Self-hosted
- Best for: Small open models and Microsoft's in-house frontier models
- Permalink: https://theaiengineering.com/platforms/model-providers#microsoft-ai

### MiniMax

MiniMax develops the MiniMax M-series language models, released as open weights, along with video, speech, and music models, and offers them through its API platform and the MiniMax Code agent.

- Website: https://www.minimax.io
- GitHub: https://github.com/MiniMax-AI
- Headquarters: Shanghai, China
- Founded: 2021
- Open source: No
- Deployment: API, Self-hosted
- Best for: Open-weight agentic LLMs plus speech and video models
- Permalink: https://theaiengineering.com/platforms/model-providers#minimax

### Mistral AI

Mistral AI releases open-weight models, many under the Apache-2.0 license, alongside commercial models, and offers them through its API, Mistral Studio, the Forge customization service, and on-premises or VPC deployments.

- Website: https://mistral.ai
- GitHub: https://github.com/mistralai
- Headquarters: Paris, France
- Founded: 2023
- Open source: No
- Deployment: API, Self-hosted
- Languages: Python, TypeScript
- Best for: Self-deployable models from a European provider
- Permalink: https://theaiengineering.com/platforms/model-providers#mistral-ai

### Moonshot AI (Kimi)

Moonshot AI develops the Kimi models, publishes open weights for its large Kimi mixture-of-experts models on Hugging Face, and serves them through the Kimi API platform for coding, agentic, and long-context work.

- Company: Moonshot AI
- Website: https://www.moonshot.ai
- GitHub: https://github.com/MoonshotAI
- Headquarters: Beijing, China
- Founded: 2023
- Open source: No
- Deployment: API, Self-hosted
- Best for: Open-weight mixture-of-experts models for agents and coding
- Permalink: https://theaiengineering.com/platforms/model-providers#moonshot-ai

### Nous Research

Nous Research is an open AI research lab that releases the Hermes family of open-weight models, often post-trained from other open base models, and maintains Hermes Agent, an MIT-licensed open-source AI agent.

- Website: https://nousresearch.com
- GitHub: https://github.com/NousResearch
- Open source: No
- Deployment: Self-hosted
- Best for: Open-weight Hermes models and agent tooling
- Permalink: https://theaiengineering.com/platforms/model-providers#nous-research

### NVIDIA Nemotron

NVIDIA's Nemotron family includes open-weight language, reasoning, speech, and multimodal models released with training datasets under the NVIDIA Nemotron Open Model License, and packaged as NIM microservices for deployment.

- Company: NVIDIA
- Website: https://www.nvidia.com/en-us/ai-data-science/foundation-models/nemotron/
- GitHub: https://github.com/NVIDIA-NeMo
- Headquarters: Santa Clara, United States
- Open source: No
- Deployment: API, Self-hosted
- Best for: Open models with published data, tuned for NVIDIA hardware
- Permalink: https://theaiengineering.com/platforms/model-providers#nvidia-nemotron

### OpenAI

OpenAI develops the GPT model family, offered through its API for text, reasoning, tool calling, image generation, audio, and realtime voice, and publishes the gpt-oss open-weight models under the Apache-2.0 license.

- Website: https://openai.com
- GitHub: https://github.com/openai
- Headquarters: San Francisco, United States
- Founded: 2015
- Open source: No
- Deployment: API, Self-hosted
- Languages: Python, TypeScript, Go, Java, C#
- Best for: One API for text, voice, and image models
- Permalink: https://theaiengineering.com/platforms/model-providers#openai

### Reflection AI

Reflection AI released Beam, a 501-billion-parameter mixture-of-experts model with weights under Apache-2.0, available through the Reflection API, the Mirror CLI, and self-hosting with SGLang, vLLM, or TokenSpeed.

- Website: https://reflection.ai
- Headquarters: New York, United States
- Founded: 2024
- Open source: No
- Deployment: API, Self-hosted
- Best for: Large Apache-licensed open weights from a US lab
- Permalink: https://theaiengineering.com/platforms/model-providers#reflection-ai

### Reka

Reka builds natively multimodal models for video, image, audio, and text, including the Spark, Edge, Flash, and Core lines, has released Reka Flash 3 weights under Apache-2.0, and now focuses on omni models for physical AI.

- Company: Reka AI
- Website: https://reka.ai
- Open source: No
- Deployment: API, Self-hosted
- Best for: Multimodal video and audio understanding
- Permalink: https://theaiengineering.com/platforms/model-providers#reka

### Sarvam AI

Sarvam AI builds models for Indian languages, including the open-weight Sarvam-30B and Sarvam-105B under Apache-2.0, plus speech-to-text, text-to-speech, and document parsing APIs on its sovereign AI platform.

- Website: https://www.sarvam.ai
- GitHub: https://github.com/sarvamai
- Headquarters: Bengaluru, India
- Founded: 2023
- Open source: No
- Deployment: API, Self-hosted
- Best for: Indian-language text and voice applications
- Permalink: https://theaiengineering.com/platforms/model-providers#sarvam-ai

### SpaceXSI (formerly xAI)

SpaceXSI, the AI company formerly named xAI, develops the Grok models and offers them through the Grok API for text, reasoning, coding, function calling, voice, and image and video generation.

xAI now operates as SpaceXSI (SpaceX Super Intelligence), while the Grok models, the Grok API, and the x.ai and docs.x.ai domains keep their names.

- Company: SpaceX
- Website: https://x.ai
- GitHub: https://github.com/xai-org
- Open source: No
- Deployment: API
- Languages: Python
- Best for: Grok models for text, code, voice, and media
- Permalink: https://theaiengineering.com/platforms/model-providers#xai

### StepFun

StepFun develops the Step language, vision, and audio models, publishes many of them as open weights under the Apache-2.0 license, and serves them through the StepFun Open Platform API.

- Website: https://www.stepfun.com
- GitHub: https://github.com/stepfun-ai
- Headquarters: Shanghai, China
- Founded: 2023
- Open source: No
- Deployment: API, Self-hosted
- Best for: Open-weight Step models and speech APIs
- Permalink: https://theaiengineering.com/platforms/model-providers#stepfun

### Tencent Hy (Hunyuan)

Tencent's Hunyuan team, whose newer models carry the Hy name, develops language, image, video, and 3D models, publishes many as open weights on Hugging Face, including the Apache-2.0 Hy4 preview, and offers them through Tencent Cloud.

- Company: Tencent
- Website: https://aistudio.tencent.com
- GitHub: https://github.com/Tencent-Hunyuan
- Headquarters: Shenzhen, China
- Open source: No
- Deployment: API, Self-hosted
- Best for: Open-weight text, image, video, and 3D models
- Permalink: https://theaiengineering.com/platforms/model-providers#tencent-hy

### Thinking Machines Lab

Thinking Machines Lab released Inkling, a multimodal mixture-of-experts model with controllable reasoning effort, as open weights under Apache-2.0, and runs Tinker, an API for fine-tuning open-weight models.

- Website: https://thinkingmachines.ai
- GitHub: https://github.com/thinking-machines-lab
- Headquarters: San Francisco, United States
- Founded: 2025
- Open source: No
- Deployment: API, Self-hosted
- Best for: Fine-tuning open-weight models through a managed API
- Permalink: https://theaiengineering.com/platforms/model-providers#thinking-machines-lab

### TII (Falcon)

The Technology Innovation Institute in Abu Dhabi develops the Falcon models, including the hybrid Falcon-H1 series and Falcon Perception vision models, and releases them as open weights under its Falcon license or Apache-2.0.

- Company: Technology Innovation Institute
- Website: https://falconllm.tii.ae
- GitHub: https://github.com/tiiuae
- Headquarters: Abu Dhabi, United Arab Emirates
- Founded: 2020
- Open source: No
- Deployment: Self-hosted
- Best for: Open-weight models from a government research institute
- Permalink: https://theaiengineering.com/platforms/model-providers#tii-falcon

### TypeSafe AI

TypeSafe AI builds Jev, a proprietary System One model served through its own API that answers typed questions about program state with structured values and calibrated probabilities instead of generated text.

- Website: https://typesafe.ai
- GitHub: https://github.com/typesafe-ai
- Headquarters: San Francisco, United States
- Open source: No
- Deployment: API
- Best for: Typed decisions inside automation workflows
- Permalink: https://theaiengineering.com/platforms/model-providers#typesafe-ai

### Upstage (Solar)

Upstage, a South Korean AI company, develops the Solar language models, including Solar Pro, Solar Mini, and the open-weight Solar Open line, alongside Document Parse and other document intelligence APIs.

- Company: Upstage
- Website: https://www.upstage.ai
- GitHub: https://github.com/UpstageAI
- Headquarters: Yongin, South Korea
- Founded: 2020
- Open source: No
- Deployment: API, Self-hosted
- Best for: Korean and Japanese models and document parsing
- Permalink: https://theaiengineering.com/platforms/model-providers#upstage

### Writer (Palmyra)

Writer is an enterprise AI platform company that trains its own Palmyra models, including the Palmyra X6 model for agentic work, offers them inside its agent platform and through an API, and has released smaller Palmyra models openly.

- Company: Writer
- Website: https://writer.com
- GitHub: https://github.com/writer
- Headquarters: San Francisco, United States
- Founded: 2020
- Open source: No
- Deployment: API
- Best for: Enterprise agents on a vendor's own models
- Permalink: https://theaiengineering.com/platforms/model-providers#writer-palmyra

### Xiaomi MiMo

Xiaomi's MiMo team develops the MiMo language, reasoning, and speech models, releases them as open weights under the MIT license, and serves them through the Xiaomi MiMo API Open Platform and the MiMo Code agent.

- Company: Xiaomi
- Website: https://mimo.xiaomi.com
- GitHub: https://github.com/XiaomiMiMo
- Headquarters: Beijing, China
- Open source: No
- Deployment: API, Self-hosted
- Best for: MIT-licensed reasoning and coding models
- Permalink: https://theaiengineering.com/platforms/model-providers#xiaomi-mimo

### Z.ai (GLM)

Z.ai, the international brand of Zhipu AI, develops the GLM language and vision models, releases many of them as open weights under the MIT license, and serves them through its API and the Z.ai assistant.

- Company: Zhipu AI
- Website: https://z.ai
- GitHub: https://github.com/zai-org
- Headquarters: Beijing, China
- Open source: No
- Deployment: API, Self-hosted
- Best for: MIT-licensed GLM models for coding and agents
- Permalink: https://theaiengineering.com/platforms/model-providers#z-ai

## How to choose

1. **Model access**: Whether models are available only through a hosted API or also as downloadable weights that run on your own infrastructure.
2. **Weight licensing**: For open-weight models, whether the license is OSI-approved, such as MIT or Apache-2.0, or a custom license with usage or revenue conditions.
3. **Capabilities and modalities**: Support for reasoning, tool calling, structured outputs, long context, and image, audio, or video input and output.
4. **Platform availability**: Whether the models are also sold through Amazon Bedrock, Google Cloud, Microsoft Foundry, or inference providers, which affects procurement and data residency.
5. **Data handling**: Retention policies, regional processing, and options for zero data retention or private deployment.
6. **SDKs and API compatibility**: Official SDK languages and whether the API accepts OpenAI-format requests, which lowers the cost of switching.

## Frequently asked questions

### What is an LLM provider?

An LLM provider is a company that trains large language models and offers them to developers, usually through an API billed per token. OpenAI, Anthropic, Google, and Mistral AI are examples. Some providers also release open weights, and many of their models are resold through cloud platforms such as Amazon Bedrock, Microsoft Foundry, and Google Cloud.

### What is an open weight model?

An open-weight model is a model whose trained parameters are published for anyone to download and run, usually on Hugging Face. Its training data and training code may stay private, and its license may add conditions. The Open Source Initiative's definition of open source AI also asks for enough information about training data to rebuild the system, so many MIT- or Apache-2.0-licensed models count only as open-weight.

### Open source vs closed source LLM: how do you choose?

Closed models served through an API require no infrastructure to operate, while open-weight models can run on your own hardware, be fine-tuned, and keep data inside your network. A common pattern combines both: a hosted proprietary model for the hardest tasks and a self-hosted open-weight model for high-volume, latency-sensitive, or privacy-restricted workloads. Evaluate both on your own tasks before committing.

### Which Chinese companies make LLMs?

Chinese model developers include DeepSeek, Alibaba's Qwen team, Moonshot AI, Z.ai (Zhipu AI), MiniMax, ByteDance Seed, Baidu, Tencent, StepFun, Xiaomi, Meituan, and Ant Group's inclusionAI. Many of them publish open weights on Hugging Face, often under MIT or Apache-2.0 licenses, and serve their models through APIs that accept OpenAI-format requests, so they can be tested alongside US models with little code change.

### How do you use multiple LLM providers in one app?

Most teams place an AI gateway or routing layer between the application and the providers. The gateway exposes one API, usually in the OpenAI format, translates requests for each provider, and adds failover, rate limits, cost tracking, and caching. Provider-agnostic SDKs are a lighter option. In both cases, keeping prompts and evaluations portable turns a model switch into a configuration change rather than a rewrite.

## Related guides

- [LLM benchmarks explained: what they measure and how to read them](https://theaiengineering.com/articles/llm-benchmarks): LLM benchmarks explained: what MMLU-Pro, GPQA, SWE-bench, LiveCodeBench, ARC-AGI, and LMArena measure, why scores saturate, and how to read a leaderboard.
- [Reasoning models: how they work and when to use them](https://theaiengineering.com/articles/reasoning-models): Reasoning models explained: how test-time compute and reinforcement learning produce chain-of-thought reasoning, when to use them, and their cost and latency.
- [What is an AI gateway? How it works and when you need one](https://theaiengineering.com/articles/what-is-an-ai-gateway): An AI gateway is a proxy between your apps and model providers. Learn how it works, its core features, AI gateway vs API gateway, and open source options.

Product details change quickly; check each vendor's site for current features, licensing and pricing.

## Cite this page

The AI Engineering Editors. "Model providers." The AI Engineering. https://theaiengineering.com/platforms/model-providers

- Title: Model providers
- Author: The AI Engineering Editors
- Publisher: The AI Engineering
- URL: https://theaiengineering.com/platforms/model-providers
- Markdown: https://theaiengineering.com/platforms/model-providers.md
