# Use BigModel.io with Continue

> Add BigModel.io models to Continue in VS Code or JetBrains with provider openai, apiBase https://bigmodel.io/v1 and a config.yaml model block, including Agent mode tool use.

Web version: https://bigmodel.io/docs/continue · Updated 2026-10-18

Continue configures models in `~/.continue/config.yaml`. Its `openai` provider accepts a custom `apiBase`, so BigModel.io models work by pointing it at `https://bigmodel.io/v1`.

| Field | Value |
|---|---|
| `provider` | `openai` |
| `apiBase` | `https://bigmodel.io/v1` |
| `apiKey` | your BigModel.io API key (`sk-...`); keep this file out of version control |
| `model` | a model ID from the [model catalog](https://bigmodel.io/models), e.g. `glm-5.3` |
| `roles` | `chat`, `edit`, `apply` (and others as needed) |
| `capabilities` | `tool_use` for Agent mode; `image_input` if the model accepts images |
| `useResponsesApi` | `false` to force Chat Completions |

## Configure config.yaml

Open `~/.continue/config.yaml` (create it if needed) and add:

```yaml
name: BigModel
version: 0.0.1
schema: v1
models:
  - name: GLM-5.3 (BigModel)
    provider: openai
    model: glm-5.3
    apiBase: https://bigmodel.io/v1
    apiKey: sk-your-bigmodel-key
    useResponsesApi: false
    roles: [chat, edit, apply]
    capabilities: [tool_use]
  - name: DeepSeek V4 Flash (BigModel)
    provider: openai
    model: deepseek-v4-flash
    apiBase: https://bigmodel.io/v1
    apiKey: sk-your-bigmodel-key
    useResponsesApi: false
    roles: [chat, edit, apply, summarize]
```

- `model` must be the exact catalog ID; `name` is only the label in the model dropdown.
- `useResponsesApi: false` keeps requests on `/v1/chat/completions`, which every text group serves. Chat Completions-only GLM and DeepSeek groups return 404 on `/v1/responses`.
- `capabilities: [tool_use]` is needed for Agent mode. Add `image_input` only for models that accept images.

> The key sits in plain text in `~/.continue/config.yaml`. Don't commit this file or copy it into a project repository.

**Autocomplete is not supported.** BigModel.io serves chat endpoints only (`/v1/chat/completions`, `/v1/responses`, `/v1/messages`); there is no legacy `/v1/completions` or fill-in-the-middle endpoint, which Continue's `autocomplete` role needs. Use BigModel.io models for the `chat`, `edit` and `apply` roles, and leave tab autocomplete on another provider or turned off.

## Verify the setup

Open the Continue chat panel, pick **GLM-5.3 (BigModel)** in the model dropdown and send "ping". You should get a reply, and the request should appear in [Usage](https://bigmodel.io/usage).

## Troubleshooting

**The model does not appear in the dropdown.** `config.yaml` has a YAML error, or Continue is using a different config. Fix the indentation and make sure Continue has your local config selected.

**404 `not_found_error`: `... /v1/responses is not available`.** Continue used the Responses API on a Chat Completions-only group. Set `useResponsesApi: false`.

**404 on every request.** `apiBase` must be `https://bigmodel.io/v1`, without `/chat/completions`.

**401 `API_KEY_REQUIRED` or `INVALID_API_KEY`.** The `apiKey` value is empty or wrong. Copy the key again from [API Keys](https://bigmodel.io/keys) and save the file.

**Agent mode never calls tools.** Add `capabilities: [tool_use]` to the model, and use a model that supports tool calling.

**503 `api_error`: `No available accounts: no available accounts supporting model: ...`.** `model` is misspelled or not served by your key's group.

**403 `INSUFFICIENT_BALANCE` or 429.** Top up in [Billing](https://bigmodel.io/purchase). A 429 `API_KEY_QUOTA_EXHAUSTED` means raise the key's quota in [API Keys](https://bigmodel.io/keys).

## Related

- [Cline](https://bigmodel.io/docs/cline) for an agent-first VS Code extension.
- [Cursor](https://bigmodel.io/docs/cursor) if you use Cursor as your editor.
- [Model IDs](https://bigmodel.io/docs/model-ids) for how IDs are written.