FeatureMatrix.md

Feature Matrix

✅ - Supported
⚠️ - Partially supported/under construction
❌ - Not supported
Empty field means the feature is unsupported by the provider.

Inference

	OpenAI	Anthropic	Google	Cohere	DeepSeek	Groq	Mistral	xAI	Custom*
Audio	✅
Chat	✅	✅	✅	✅	❌	✅	❌	❌	✅
Embeddings	✅		❌	❌			❌	❌
Fine-Tuning	✅		❌				❌
Batch	❌	❌	❌	❌		❌	❌
Files	✅		✅				❌
Uploads	❌
Images	✅		❌
Models	✅	❌	❌	❌	❌	❌	❌	❌	✅
Moderation	✅
Tokenize		❌		❌

*Custom means any OpenAI compatible provider, such as Azure OpenAI, Ollama, KoboldCpp, etc.

OpenAI Specific

Assistants	Threads	Messages	Runs	Run steps	Vector stores	Vector store files	Vector store file batches	Realtime
✅	✅️	✅️	✅️	✅️	✅	✅	✅	❌

Google Specific

Caching*	Semantic Retrieval
✅	❌

*Other providers expose caching as part of /chat, or don't offer the feature.

Cohere Specific

Rerank	Embed Jobs*	Classify*	Datasets*	Connectors*
❌	❌	❌	❌	❌

*/v1 APIs, future support is unsure and probably won't be implemented.

Mistral Specific

FIM	Agents	Classifiers
❌	❌	❌

Misc

Some providers offer /admin APIs, LlmTornado won't ship support for these, as there is little to no overlap among them. We recommend using autogenerated APIs for the consumption of these services.