Skip to main content
Docs
Providers

Providers

Learn about the AI providers available through Elyxir and their model offerings.

Elyxir provides access to models from 9 major AI providers through a single unified API. Each provider has unique strengths and model offerings.

Supported Providers

OpenAI

GPT-4o, GPT-4, GPT-3.5, DALL-E, Whisper

Anthropic

Claude 3.5 Sonnet, Claude 3 Opus/Haiku

Google

Gemini Pro, Gemini Flash

Meta

Llama 3.2, Llama 3.1

OpenAI

Industry-leading AI models with broad capabilities.

Available Models

ModelBest ForContext
GPT-4oGeneral purpose, fast128K
GPT-4o-miniCost-effective tasks128K
GPT-4 TurboComplex reasoning128K
GPT-3.5 TurboSimple, fast tasks16K
text-embedding-3-largeHigh-quality embeddings8K
text-embedding-3-smallCost-effective embeddings8K

Strengths

  • Excellent general-purpose performance
  • Fast response times
  • Strong vision capabilities
  • Reliable and well-documented

Anthropic

Known for helpful, harmless, and honest AI.

Available Models

ModelBest ForContext
Claude 3.5 SonnetCoding, analysis200K
Claude 3 OpusMost capable200K
Claude 3 SonnetBalanced200K
Claude 3 HaikuFast, affordable200K

Strengths

  • Exceptional coding abilities
  • Very long context window
  • Strong reasoning
  • Detailed, thoughtful responses

Google

Multimodal AI with broad capabilities.

Available Models

ModelBest ForContext
Gemini 1.5 ProComplex tasks1M
Gemini 1.5 FlashFast responses1M
Gemini Pro VisionImage analysis32K

Strengths

  • Extremely large context windows
  • Strong multimodal capabilities
  • Good price-performance ratio
  • Fast response times

Meta

Open-weight models for flexibility.

Available Models

ModelBest ForContext
Llama 3.2 90BComplex tasks128K
Llama 3.2 11BBalanced128K
Llama 3.1 405BMaximum capability128K
Llama 3.1 70BGood balance128K

Strengths

  • Open weights (can self-host)
  • No usage restrictions
  • Strong coding abilities
  • Good multilingual support

Cohere

Enterprise-focused with strong RAG support.

Available Models

ModelBest ForContext
Command R+RAG, enterprise128K
Command RGeneral tasks128K
Embed v3High-quality embeddings512

Strengths

  • Excellent for RAG applications
  • Strong citation capabilities
  • Enterprise features
  • Multilingual embeddings

Perplexity

Real-time information with web search.

Available Models

ModelBest ForContext
Sonar LargeDeep research28K
Sonar MediumGeneral search28K

Strengths

  • Native web search integration
  • Real-time information
  • Source citations
  • Current events knowledge

xAI

Latest entrant with strong capabilities.

Available Models

ModelBest ForContext
GrokGeneral conversation128K

Strengths

  • Real-time information
  • Witty responses
  • Broad knowledge base

ElevenLabs

Best-in-class text-to-speech.

Available Models

ModelBest For
Eleven Multilingual v2High-quality TTS
Eleven Turbo v2Fast TTS

Strengths

  • Natural-sounding voices
  • Multiple languages
  • Voice cloning
  • Fast generation

Choosing a Provider

For General Use

  • Best quality: OpenAI GPT-4o or Anthropic Claude 3.5 Sonnet
  • Best value: OpenAI GPT-4o-mini or Meta Llama 3.2

For Coding

  • Best: Anthropic Claude 3.5 Sonnet
  • Alternative: OpenAI GPT-4o

For Research

  • Best: Perplexity Sonar (with web search)
  • Alternative: Google Gemini (large context)

For Embeddings

  • Best quality: OpenAI text-embedding-3-large
  • Best value: Cohere Embed v3
Providers | elyxir