Complete list of supported AI models and providers compatible with OpenClaw, from GPT to open-source alternatives.
What AI Models Can I Use with OpenClaw?
OpenClaw is model-agnostic by design — it works with virtually any large language model (LLM) available through an API or running locally. This flexibility lets you choose the best model for your needs, budget, and privacy requirements.
# Switch to Anthropic Claude
model:
provider: anthropic
model: claude-3.5-sonnet
api_key: ${ANTHROPIC_API_KEY}
# Switch to a local model via Ollama
model:
provider: ollama
model: llama3.1:70b
endpoint: http://localhost:11434
# No API key needed!
Model Comparison for OpenClaw Use
Model
Best For
Speed
Cost
Quality
GPT-4o
General assistant, complex reasoning
Fast
$$
★★★★★
Claude 3.5 Sonnet
Long documents, nuanced writing
Fast
$$
★★★★★
Gemini 2.5 Flash
Quick tasks, high volume
Very fast
$
★★★★
Llama 3.1 70B
Privacy-first, offline use
Medium
Free
★★★★
Llama 3.1 8B
Low-resource hardware
Fast
Free
★★★
Mistral 7B
Lightweight, fast responses
Very fast
Free
★★★
Mixtral 8x7B
Balanced local option
Medium
Free
★★★★
GPT-3.5 Turbo
Budget cloud option
Very fast
¢
★★★
Using Local Models with Ollama
Ollama is the recommended way to run local models with OpenClaw:
# Install Ollama
curl -fsSL https://ollama.ai/install.sh | sh
# Download a model
ollama pull llama3.1:70b
# Verify it's running
ollama list
# Configure OpenClaw
openclaw config set model.provider ollama
openclaw config set model.model llama3.1:70b
Using Multiple Models
OpenClaw supports model routing — using different models for different tasks:
model:
default:
provider: ollama
model: llama3.1:8b # Fast, free for simple tasks
complex:
provider: openai
model: gpt-4o # Heavy lifting
trigger: "complex_reasoning"
coding:
provider: anthropic
model: claude-3.5-sonnet # Best for code
trigger: "code_generation"
Hardware Requirements for Local Models
Model Size
RAM Required
GPU VRAM
Example Models
7B parameters
8 GB
6 GB
Mistral 7B, Llama 3.1 8B
13B parameters
16 GB
10 GB
CodeLlama 13B
34B parameters
32 GB
24 GB
CodeLlama 34B
70B parameters
64 GB
48 GB
Llama 3.1 70B
Tip: You can run 7B models on most modern laptops. For 70B models, consider a desktop with a high-end GPU or use quantized (Q4/Q5) versions to reduce memory requirements.
Switching Models via CLI
# Check current model
openclaw config get model
# Switch provider
openclaw config set model.provider anthropic
openclaw config set model.model claude-3.5-sonnet
openclaw config set model.api_key $ANTHROPIC_API_KEY
# Test the new model
openclaw chat "Hello, which model are you?"
# Switch to local
openclaw config set model.provider ollama
openclaw config set model.model llama3.1:8b
Troubleshooting Model Issues
Problem
Solution
"API key invalid"
Verify the key in your provider's dashboard
"Model not found"
Check the exact model name (case-sensitive)
Slow responses (local)
Use a smaller model or enable GPU acceleration
"Context too large"
Reduce max_tokens or switch to a model with larger context
Small Language Models On-Device — The Quiet Revolution of 2026 — Everyone is watching GPT-5 and Claude 4.5, but the real shift in 2026 is happening on the device. Phi-4, Gemma 3, and Llama 3.3-3B now run on laptops and phones at GPT-3.5 quality. Here is what that means for the apps you build.