Skip to main content
Anyware supports running Claude Code with dozens of different models via OpenRouter, giving you access to Claude, GPT-4, Gemini, Llama, Mistral, and many more.

​
Quick Start

Use the --model flag to specify a model: 
anyware --model openai/gpt-4o

​
Model Format

Models are specified in the format provider/model-name: 
provider/model-name
Examples:
  • anthropic/claude-sonnet-4
  • openai/gpt-4o
  • google/gemini-pro
  • meta-llama/llama-3-70b

​
Available Models

​
Anthropic (Claude)

ModelID
Claude Opus 4anthropic/claude-opus-4
Claude Sonnet 4anthropic/claude-sonnet-4
Claude 3.5 Sonnetanthropic/claude-3.5-sonnet
Claude 3 Haikuanthropic/claude-3-haiku

​
OpenAI

ModelID
GPT-4oopenai/gpt-4o
GPT-4o Miniopenai/gpt-4o-mini
GPT-4 Turboopenai/gpt-4-turbo
o1 Previewopenai/o1-preview
o1 Miniopenai/o1-mini

​
Google

ModelID
Gemini Progoogle/gemini-pro
Gemini Pro Visiongoogle/gemini-pro-vision
Gemini 1.5 Progoogle/gemini-1.5-pro

​
Meta (Llama)

ModelID
Llama 3 70Bmeta-llama/llama-3-70b
Llama 3 8Bmeta-llama/llama-3-8b
Llama 2 70Bmeta-llama/llama-2-70b-chat

​
Mistral

ModelID
Mistral Largemistralai/mistral-large
Mistral Mediummistralai/mistral-medium
Mixtral 8x7Bmistralai/mixtral-8x7b

​
Setting a Default Model

​
Environment Variable

Set a default model using an environment variable: 
export ANYWARE_MODEL=openai/gpt-4o
anyware
Add to your shell profile for persistence: 
# ~/.bashrc or ~/.zshrc
export ANYWARE_MODEL=openai/gpt-4o

​
Per-Session Override

The --model flag always overrides the default: 
# Uses default model from ANYWARE_MODEL
anyware

# Overrides with specific model
anyware --model anthropic/claude-opus-4

​
Model Considerations

​
Compatibility

Not all models have the same capabilities:
FeatureClaudeGPT-4GeminiLlama
Code generationExcellentExcellentGoodGood
Context window200K128K1M4K-128K
Tool useYesYesYesLimited
VisionYesYesYesNo

​
Performance

Different models have different latencies:
  • Fastest: Claude Haiku, GPT-4o Mini, Llama 3 8B
  • Balanced: GPT-4o, Claude Sonnet 4
  • Highest Quality: Claude Opus 4, GPT-4 Turbo, o1

​
Best Practices

​
Match Model to Task

# Quick tasks, low cost
anyware --model anthropic/claude-3-haiku

# Complex coding tasks
anyware --model anthropic/claude-opus-4

# General purpose, good balance
anyware --model openai/gpt-4o

​
Use Shell Aliases

Create shell aliases for common configurations: 
# Add to ~/.bashrc or ~/.zshrc
alias anyware-fast='anyware --model anthropic/claude-3-haiku'
alias anyware-smart='anyware --model anthropic/claude-opus-4'
alias anyware-gpt='anyware --model openai/gpt-4o'

​
Monitor Usage

Track your token usage in the web dashboard at anyware.run to optimize costs.

​
Troubleshooting

​
Model Not Working as Expected

Different models have different behaviors. If a model isn’t performing well:
  1. Check the model’s known limitations
  2. Try adjusting your prompts for that model
  3. Fall back to Claude for complex coding tasks

​
Invalid Model Error

If you see an error about an invalid model:
  1. Check the model ID format (provider/model-name)
  2. Verify the model is available on OpenRouter
  3. Try a known working model like anthropic/claude-sonnet-4

​
See Also