Available Models
Every Hosted Claws plan includes AI models — no API keys required. Models are hosted on DigitalOcean Gradient and billed as part of your plan.
Model Catalog
Standard Tier ($25/mo)
| Model | Parameters | Context | Best For |
|---|---|---|---|
| GPT-OSS 20B | 21B (3.6B active) | 128k | Daily tasks, chat, coding — fast and capable |
| GPT-OSS 120B | 120B (5.1B active) | 128k | Complex reasoning, longer tasks |
| GPT-5 nano | — | — | Quick responses, simple queries |
All three models support tool calling and chain-of-thought reasoning.
Pro Tier ($50/mo)
Everything in Standard, plus:
| Model | Parameters | Context | Best For |
|---|---|---|---|
| GPT-5 mini | — | — | Stronger reasoning, multi-step tasks |
| Llama 3.3 70B | 70B | 128k | Open-weight alternative, strong at code |
Enterprise Tier ($200/mo)
Everything in Pro, plus:
| Model | Parameters | Context | Best For |
|---|---|---|---|
| GPT-5.2 | — | — | Frontier reasoning and coding |
| GPT-5.3 Codex | — | — | Specialized code generation |
| Claude Sonnet 4.6 | — | 200k | Nuanced writing, analysis, long context |
| Claude Opus 4.6 | — | 200k | Most capable Claude — complex tasks |
How Models Are Selected
Your OpenClaw assistant automatically selects the best model for each task based on complexity. You can also specify a model preference in your assistant's configuration.
About GPT-OSS
GPT-OSS models are OpenAI's open-weight models released under the Apache 2.0 license. They use a Mixture-of-Experts (MoE) architecture — only a fraction of parameters are active per token, making them fast and efficient while maintaining strong benchmark scores.
The 20B model matches o3-mini on most benchmarks, making it an exceptional value for the Standard tier.