Skip to main content
Docs
Model Discovery

Model Discovery

Learn how to find and filter models in the Elyxir Model Hub.

The Model Hub makes it easy to find the right AI model for your use case. With 2,000+ models from 9 providers, effective discovery tools are essential.

Type in the search bar to find models:

  • By name: "gpt-4", "claude", "llama"
  • By provider: "openai", "anthropic"
  • By capability: "vision", "embeddings"

Search Tips

  • Use partial names: "gpt" finds all GPT models
  • Search is case-insensitive
  • Multiple terms narrow results

Filters

By Provider

Filter to show models from specific providers:

  • OpenAI
  • Anthropic
  • Google
  • Meta
  • Cohere
  • Perplexity
  • xAI
  • ElevenLabs
  • Ollama (self-hosted)

By Capability

Filter by what the model can do:

CapabilityDescription
ChatConversational responses
CodeProgramming and code generation
VisionImage understanding and analysis
EmbeddingsText vectorization for search/RAG
AudioSpeech-to-text, text-to-speech
MultimodalMultiple input/output types

By Context Window

Filter by maximum context size:

  • 4K tokens: Basic conversations
  • 8K tokens: Short documents
  • 32K tokens: Long documents
  • 128K+ tokens: Very long context

By Price

Filter by cost tier:

  • Free: Experimental/limited models
  • Budget: < $1 per 1M tokens
  • Standard: $1-10 per 1M tokens
  • Premium: > $10 per 1M tokens

Sorting

Sort the model list by:

Sort OptionDescription
PopularMost used models
NewestRecently added
NameAlphabetical
Price (Low)Cheapest first
Price (High)Most expensive first
Context SizeLargest context first

Model Cards

Each model card shows at a glance:

┌─────────────────────────────────┐
│ Provider Logo                   │
│ Model Name                      │
│ ─────────────────────────────── │
│ Capabilities: 💬 👁️ 💻          │
│ Context: 128K tokens            │
│ Price: $0.005 / 1K tokens       │
│                   [Try] [More]  │
└─────────────────────────────────┘

Finding the Right Model

For Chat Applications

Best for quality: GPT-4o, Claude 3.5 Sonnet Best for speed: GPT-4o-mini, Claude 3 Haiku Best for cost: Llama 3.2, Mistral

For Code Generation

Best overall: Claude 3.5 Sonnet Good balance: GPT-4o Open source: CodeLlama, DeepSeek Coder

For Vision Tasks

Best quality: GPT-4V, Claude 3 Opus Good balance: GPT-4o, Claude 3.5 Sonnet Fast: Gemini Flash

For Embeddings

Best quality: text-embedding-3-large Good balance: text-embedding-3-small Alternative: Cohere Embed v3

Saved Filters

Save frequently used filter combinations:

  1. Apply your filters
  2. Click Save Filter
  3. Name your filter
  4. Access from the Saved dropdown
Model Discovery | elyxir