[model]
LLM provider, model, context window, and reasoning settings for this project.
Required at the top level of seal.toml.
Provider-specific model identifier to request.
provider
Section titled “provider”Provider backend to use for LLM requests, such as anthropic or openai.
context_length
Section titled “context_length”Context-window override for this model. When omitted, Seal uses provider metadata.
effort
Section titled “effort”Reasoning effort to request from providers that support effort levels. When omitted, Seal chooses a model-appropriate default.
| Value | Meaning |
|---|---|
"low" | — |
"medium" | — |
"high" | — |
"xhigh" | Between High and Max. Anthropic’s recommended starting point for coding / agentic workloads on Opus 4.7. |
"max" | Anthropic’s highest effort level. Opus-4.6+ only; other models return a 400 when this value is sent. |
headers
Section titled “headers”Extra HTTP headers merged into every LLM request.
max_tokens
Section titled “max_tokens”Output-token cap to send to the provider as max_tokens. When omitted, Seal uses provider metadata to pick the model’s advertised ceiling.
thinking
Section titled “thinking”Extended thinking mode for LLM requests.
| Value | Meaning |
|---|---|
"adaptive" | Model decides how much to think (default). |
"manual" | Explicit token budget for thinking (requires thinking_budget). |
"disabled" | No thinking. |
thinking_budget
Section titled “thinking_budget”Token budget for manual thinking mode. Required when thinking = "manual".