bots.so — The AI Inference Model Index
Extracto
Discover and compare AI inference models across providers. Real-time pricing, availability, and performance data for GPT-4, Claude, Llama, Gemini, and 50+ more models.
Resumen
Resumen Principal
El panorama actual de proveedores de inferencia de Modelos de Lenguaje Grandes (LLM) es dinámico y diverso, abarcando 42 entidades clasificadas en tiers como "major" y "community". Un hallazgo central es la prevalencia de la compatibilidad con la API de OpenAI entre la mayoría de los actores, un factor clave que
Contenido
42 Providers Tracked
All LLM Inference Providers
Live tracked count. Showing 42 profiled providers across flagship, major, and community tiers.
Flagship Providers
Major Providers
Live
- Together AI
- Run leading open-source models with one API
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Frontier Deck Frontier Deck means this provider carries an extensive lineup across many model categories.
North America This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Groq
- Fastest inference on custom LPU hardware
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Global This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- NVIDIA
- AI infrastructure and inference at scale
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Frontier Deck Frontier Deck means this provider carries an extensive lineup across many model categories.
US + Global This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Ollama
- Run AI models anywhere
- Models Hosted
- API Style
- Native + OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Undisclosed (Contacted 16 Feb 2026) This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Fireworks AI
- High-performance inference at scale
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Expanded Deck Expanded Deck means this provider offers a broad lineup that covers many common model families.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Mistral AI
- European AI lab pushing efficiency frontiers
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
EU (France) This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Cerebras
- Wafer-scale AI inference at record speeds
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US + Canada + EU This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Baseten
- Model APIs for production AI
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Replicate
- Run AI models with a cloud API
- Models Hosted
- API Style
- Replicate API
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Cohere
- Enterprise AI with RAG specialization
- Models Hosted
- API Style
- Cohere API
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US + Global This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Cloudflare Workers AI
- AI at the edge, globally distributed
- Models Hosted
- API Style
- REST API
- Compute Location
Expanded Deck Expanded Deck means this provider offers a broad lineup that covers many common model families.
Global (Edge) This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Crusoe
- The world's favorite AI cloud
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US (Wyoming, Texas) This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- DeepInfra
- Simple, scalable AI inference
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Frontier Deck Frontier Deck means this provider carries an extensive lineup across many model categories.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Nebius
- The ultimate cloud for AI explorers
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
EU + US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Hyperstack
- European GPU cloud powered by renewable energy
- Models Hosted
- API Style
- Custom (api_key header)
- Compute Location
Expanded Deck Expanded Deck means this provider offers a broad lineup that covers many common model families.
Global This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Parasail
- No limits. No contracts. Priced right.
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US + Global This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Scaleway
- Managed Inference with European sovereignty
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
EU (France) This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- io.net
- Decentralized GPU Cloud
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Decentralized This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- OVHcloud
- AI Endpoints with European data sovereignty
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Expanded Deck Expanded Deck means this provider offers a broad lineup that covers many common model families.
Global (46 DCs) This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- SiliconFlow
- AI Infrastructure for LLMs & Multimodal
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Asia (China) This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Weights & Biases Inference
- Hosted inference endpoints for open models
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- IONOS
- European cloud provider with hosted AI inference
- Models Hosted
- API Style
- Managed AI Model Hub
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
EU + US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- DigitalOcean
- Gradient AI platform at the edge
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Global This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Community Providers
Live
- Hyperbolic
- Decentralized GPU network for AI inference
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- SambaNova
- Enterprise AI with custom RDU chips
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Featherless
- Serverless inference for open-source models
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Chutes
- Serverless AI Compute
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- GMI Cloud
- GPU Cloud for Scalable AI
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Taiwan + Global This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- CanopyWave
- Open-source AI inference platform
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Asia-Pacific This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Novita AI
- Open-source LLM API
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Expanded Deck Expanded Deck means this provider offers a broad lineup that covers many common model families.
Asia + Global This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Nscale
- The Hyperscaler Engineered for AI
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Expanded Deck Expanded Deck means this provider offers a broad lineup that covers many common model families.
EU + Global This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Mancer
- Creative AI for roleplay and storytelling
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Public AI
- AI as public infrastructure
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Live
- Venice AI
- The Only Private AI API
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Decentralized This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Down
- Lilypad Network
- Decentralized AI inference
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Decentralized This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Down
- Google (Legacy Key)
- Legacy key preserved for historical provider continuity
- Models Hosted
- API Style
- Google Cloud / AI Studio
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Global (GCP) This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Down
- WaveSpeed
- Platform for faster AI image and video generation
- Models Hosted
- API Style
- Platform API
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
US + Europe This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Down
- Z.ai
- Developer API platform for Z.ai models
- Models Hosted
- API Style
- Model API endpoints
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Australia + Germany This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Soon
- Inceptron
- Compiler-Optimized AI Inference
- Models Hosted
- API Style
- OpenAI-compatible
- Compute Location
Curated Deck Curated Deck means this provider runs a focused lineup of models, chosen for specific use cases.
Europe This is our best guess at where the compute is powered and provided from for this provider, based on the data we could find. May be incorrect.
Newsletter
Get the signal, skip the noise.
Weekly digest of new models and provider updates across 41+ compute providers. Curated for AI builders who ship.
New model releases
Capability updates
Provider status
Fuente: bots.so