Providers
Meet the model developers and the platforms that serve their APIs.
Showing 55providers. Region refers to company origin, not API hosting. Organizations with both in-house models and platform services appear in both groups.
ALIBABA CLOUD
ALIBABA CLOUDQwen and multimodal model services. Beijing, Singapore, and other regions have separate endpoints, keys, and model catalogs.
BAICHUAN
BAICHUANDeveloper of Baichuan language models. Verify open-weight and commercial API versions, licenses, and service scope separately.
BAIDU
BAIDUDeveloper of ERNIE, with in-house and third-party model development and deployment through Qianfan.
HUAWEI CLOUD
HUAWEI CLOUDDeveloper of Pangu and provider of the ModelArts Studio platform. Deployed Pangu and third-party models can be called through Pangu V1 or OpenAI-compatible V2 APIs; model IDs, endpoints, regions, and account access depend on the deployed service.
VOLCENGINE
VOLCENGINEAn API platform for Doubao and other models. Distinguish model IDs from inference endpoint IDs for each product.
STEPFUN
STEPFUNLanguage reasoning, vision, speech, and multimodal generation. China and international platforms are treated as separate channels.
IFLYTEK
IFLYTEKDeveloper of Spark language models and speech technology. Verify WebSocket and HTTP APIs separately.
KLING AI
KLING AIDeveloper of Kling video and Kolors image models. Creative products and developer platforms are managed separately.
01.AI
01.AIDeveloper of Yi. Verify licensing and availability separately for model weights and commercial services.
ANT LING
ANT LINGBailing and Ling models. API capabilities depend on official open documentation and platforms.
MEITUAN LONGCAT
MEITUAN LONGCATDeveloper of LongCat. Open-weight releases and API model IDs are recorded separately.
MODEL BEST
MODEL BESTDeveloper of MiniCPM, focused on efficient language and multimodal models. Open models do not imply a unified commercial API.
SENSENOVA
SENSENOVAModels for language understanding, reasoning, and multimodal tasks, with APIs on its open platform.
VIDU
VIDUDeveloper of video models for text-to-video, image-to-video, and reference-based generation.
TENCENT HUNYUAN
TENCENT HUNYUANDeveloper of Hunyuan models for text and multimodal tasks. Basic listings come from Tencent TokenHub; channels and regions are verified separately.
XIAOMI MIMO
XIAOMI MIMODeveloper of MiMo models. Check the open platform for model listings, quotas, and API access.
MOONSHOT AI
MOONSHOT AIDeveloper of Kimi models for long text, reasoning, coding, and multimodal tasks.
ZHIPU Z.AI
ZHIPU Z.AIDeveloper of GLM language and multimodal models. Verify the China platform and international services separately.
AI21 Labs
AI21 LABSDeveloper of Jamba models for enterprise language tasks and long-context processing.
Amazon / AWS
AMAZON AWSDeveloper of Nova and Titan, with additional providers available through Bedrock. Model access and regions are managed separately.
Anthropic
ANTHROPICDeveloper of Claude models for text, coding, visual understanding, and tool use.
AssemblyAI
ASSEMBLYAIAudio transcription, streaming speech recognition, and audio understanding services.
Black Forest Labs
BLACK FOREST LABSDeveloper of FLUX image models. Official services and hosted channels such as fal and Replicate are indexed separately.
Cohere
COHEREEnterprise language models, retrieval-augmented generation, embeddings, and reranking.
Deepgram
DEEPGRAMModels and APIs for speech recognition, synthesis, and real-time audio.
DeepSeek
DEEPSEEKOfficial APIs for chat, coding, and reasoning. Available models depend on current documentation and account permissions.
ElevenLabs
ELEVENLABSSpeech synthesis, voices, and recognition models supporting multilingual audio workflows.
Google Gemini
GOOGLEOfficial Gemini APIs. Interactions and generateContent use different model locations and input structures; follow the selected integration.
IBM watsonx
IBMDeveloper of Granite. The watsonx platform provides enterprise model inference and management.
Luma AI
LUMA AIImage and video generation technology. Use the respective documentation for developer APIs and creative products.
Meta
METADeveloper of Llama and other models. Check licensing and providers separately for open weights, self-hosting, and cloud APIs.
Microsoft Azure
MICROSOFTDevelops its own models and serves others through Azure. Azure deployment names differ from original model IDs.
MiniMax
MINIMAXText, speech, and multimodal models. Configure each API protocol and region according to its documentation.
Mistral AI
MISTRAL AIDeveloper of models for language, coding, vision, speech, and document understanding.
NVIDIA
NVIDIADevelops Nemotron and offers its own and third-party models through NIM and API catalogs.
OpenAI
OPENAIOriginal text, image, and audio model services. Responses and other APIs use their own parameter contracts.
Perplexity
PERPLEXITYSonar models for web search and information retrieval. Distinguish search citations from ordinary generated output.
Runway
RUNWAYVideo and visual generation models. Verify capabilities separately for creative web tools and developer APIs.
Stability AI
STABILITY AIImage generation, image editing, and audio models. Licenses and APIs are managed per product.
xAI
XAIDeveloper of Grok models for chat, reasoning, and multimodal tasks.
SILICONFLOW
SILICONFLOWA multi-model inference platform in China. Model developers and SiliconFlow service channels are labeled separately.
SCNet · National Supercomputing Internet
SCNet · National Supercomputing InternetSCNet is a China-based multi-model API platform operated by Sugon Intelligent Computing. It offers DeepSeek, Qwen, GLM, Kimi and MiniMax models with Token Plan subscriptions and OpenAI / Anthropic-compatible interfaces. Model availability, pricing and quotas follow each service’s official terms.
JD CLOUD / JOYAGENT
JD CLOUD JOYAGENTJD Cloud's JoyAgent model platform provides one API key, an OpenAI-compatible interface, and unified billing for text, reasoning, vision, embedding, reranking, and image models. Models listed there use the JD Cloud channel; model IDs, prices, free quotas, and access rights depend on each model detail page.
MODELSCOPE
MODELSCOPEA model community and inference platform. Verify licenses, developers, and serving channels separately.
QINIU CLOUD
QINIU CLOUDA cloud platform for multiple model APIs. Its model catalog and prices apply to the Qiniu channel.
Cerebras
CEREBRASModel inference using specialized compute systems. Actual models and capabilities depend on the service catalog.
Cloudflare Workers AI
CLOUDFLAREMulti-provider model inference on Cloudflare. Bindings, regions, and billing follow platform-specific rules.
DeepInfra
DEEPINFRAHosting and inference for open language and multimodal models.
fal
FAL.AIA platform for image, video, and audio models. Platform APIs are separate from each model developer's official interface.
Fireworks AI
FIREWORKS AIInference and deployment for multiple open models. Platform parameters are independent of original provider APIs.
Groq
GROQInference acceleration for open models. The hardware and service provider is distinct from the model developer.
Hugging Face
HUGGING FACEA model community and inference platform. Check the model developer, license, and actual inference provider separately.
OpenRouter
OPENROUTERAn API aggregation platform. Catalog entries and prices apply to this channel, not necessarily the original provider's API or pricing.
Replicate
REPLICATEThird-party inference organized by model and version. Input schemas are tied to a specific version.
Together AI
TOGETHER AIInference, dedicated deployment, and fine-tuning for open models. These catalog entries apply to the Together channel.