OpenCode Console lists Hy4 flash Free as free for a limited time, with model ID hy4-flash-free; in OpenCode, select it as opencode/hy4-flash-free. This free access applies to OpenCode Console and does not mean calls through Tencent Cloud TokenHub or OpenRouter are free.
Suitable tasks: Trying Hy4 flash Free offered by Console in OpenCode.
Unsuitable tasks: Applying OpenCode's limited-time free access to another client or another provider's API.
Applicable model version: Hy4 flash Free, model ID hy4-flash-free.
Applicable client, agent, or API: OpenCode; the table lists the OpenAI-compatible Chat Completions endpoint https://opencode.ai/inference/openai/v1/chat/completions.
Recommended reasoning tier and parameters: Not specified.
First, sign in to OpenCode Console and follow the page instructions to get an API key. In the OpenCode TUI, run /connect, select the OpenCode pay-as-you-go provider, and paste the key. Then run /models and select Hy4 flash Free.
You can also set it as the default model in opencode.jsonc:
{
"$schema": "https://opencode.ai/config.json",
"model": "opencode/hy4-flash-free"
}The configuration above uses OpenCode's built-in provider. The official table also lists the Chat Completions endpoint for this model. The model ID and endpoint are:
model: hy4-flash-free
endpoint: https://opencode.ai/inference/openai/v1/chat/completionsThe official page says free-tier chat models must be used through OpenCode; other clients should use a paid model or OpenCode Go. Therefore, the endpoint in the table does not mean this free model can be called directly from any client.
Sign in to OpenCode Console and get an API key. The Console's general setup instructions also mention adding billing information and credits; the page does not clearly say whether using a limited-time free model requires a positive balance.
In the OpenCode TUI, run /connect, select the OpenCode pay-as-you-go provider, and enter the key.
Run /models and select Hy4 flash Free, or set opencode/hy4-flash-free in the model field of the OpenCode configuration.
Before sending a request, check the current model list in Console to confirm that free access is still available.
The model table names the model Hy4 flash Free and gives its exact ID as hy4-flash-free.
The table lists the endpoint https://opencode.ai/inference/openai/v1/chat/completions and the AI SDK package @ai-sdk/openai-compatible.
The page says free-tier chat models are limited to OpenCode; other clients can use paid models or OpenCode Go.
The page says Hy4 flash Free is free on OpenCode for a limited time, and that its provider follows a zero-retention policy and does not use user data for model training.
The page's general setup instructions require signing in to Console, adding billing information and credits, and getting an API key; it does not separately explain whether the free model requires a payment method or account balance.
“Free for a limited time” is the current service condition for this model in OpenCode Console; eligibility and availability may change. Do not generalize it to free credits, pricing, or access eligibility on Tencent Cloud TokenHub or OpenRouter.
The OpenCode page states that the provider for Hy4 flash Free has zero retention and does not use data for training. This does not establish that Tencent TokenHub, OpenRouter, or other providers follow the same data policy.
The page does not specify Hy4 flash Free's rate limits, request limits, context length, regional availability, when billing starts, or reasoning-parameter limits. Do not infer these conditions from Hy4 parameters at other providers.
The OpenCode Console page says its models are hosted in the United States and the European Union; the Hy4 entry does not list a more specific deployment region.
This note records official setup instructions; no request was sent and account eligibility for free access was not verified.
After signing in to OpenCode Console and getting a key, connect and select the model using the TUI steps above. Because the page says free chat models must be used through OpenCode, this note does not present the OpenAI-compatible endpoint in that entry as instructions for free calls from other clients. The model table and related notes were viewable without signing in when collected.
Hy4 Flash