Hmbown/CodeWhale · error · anyhow::Error

model `{model}` is available from configured provider route(

Error message

model `{model}` is available from configured provider route(s): {providers}. Pass `--provider <provider>` with `--model {model}` to choose one explicitly. In the TUI, use `/provider`, `/model`, or `/setup` to resolve the route before sending.

What it means

Before a one-shot run with an explicit --model, the router computes which configured providers can serve that model (explicit_route_candidate_providers); if the active provider is not among them, it refuses instead of sending a request that would fail upstream. The message lists the candidate routes and names both remedies: --provider on the CLI, /provider //model //setup in the TUI.

Source

Thrown at crates/tui/src/lib.rs:10070

            );
            return Ok(CliAutoRoute {
                provider: selection.provider,
                model: selection.model,
                reasoning_effort: selection.reasoning_effort,
                auto_controls_reasoning,
                auto_model: false,
            });
        }

        let candidate_providers = model_routing::explicit_route_candidate_providers(config, model);
        if !candidate_providers.is_empty() && !candidate_providers.contains(&config.api_provider())
        {
            let providers = candidate_providers
                .iter()
                .map(|provider| provider.as_str())
                .collect::<Vec<_>>()
                .join(", ");
            bail!(
                "model `{model}` is available from configured provider route(s): {providers}. \
                 Pass `--provider <provider>` with `--model {model}` to choose one explicitly. \
                 In the TUI, use `/provider`, `/model`, or `/setup` to resolve the route before sending."
            );
        }

        // When --model is not `auto`, fall back to the reasoning_effort
        // declared in the user's config.toml. The previous hard-coded `None`
        // silently dropped the user's setting on every non-auto-route exec
        // call, which (for example) prevented vllm + Qwen3 users from
        // disabling thinking via `reasoning_effort = "off"` and caused
        // 30+ second SSE idle timeouts on trivial prompts.
        let reasoning_effort = config
            .reasoning_effort()
            .map(crate::tui::app::ReasoningEffort::from_setting);
        Ok(CliAutoRoute {
            provider: config.api_provider(),
            model: model.to_string(),

View on GitHub (pinned to 0c42157ee5)

Solutions

1. Pass `--provider <one-of-the-listed>` together with --model
2. Or switch to a --model your current provider serves
3. In the TUI, resolve the route via /provider, /model, or /setup
4. Align the route config so your default provider serves the models you use daily

Example fix

# before
codewhale --model some-model -p "hi"
# model `some-model` is available from configured provider route(s): openrouter, ...

# after
codewhale --provider openrouter --model some-model -p "hi"

# in automation, always pin the provider when pinning the model
codewhale --provider "$PROVIDER" --model "$MODEL" ...   # both or neither

When it happens

Trigger: Passing --model X when X is routed only through other configured providers -- e.g., the model id is served by an openrouter/custom route while the current api_provider points elsewhere.

Common situations: The same model id available from several providers; the default provider was switched in config while old --model habits remained; commands copied from a teammate whose route config differs.

Related errors

• max_chars must be > 0
• context_window must be greater than 0
• custom provider '{provider_id}' must set [providers.{provide
• {name}.context_window must be greater than 0
• Invalid provider '{provider}': expected {}.

AI-assisted analysis of Hmbown/CodeWhale@0c42157ee5 (2026-08-20). Data as JSON: /api/errors/cdafd8782ca7188e. Report an issue: GitHub.