Command Code's Hy4 Flash model ID is tencent/hy4-flash. The FAQ says it is available on the Go plan and above; the page provides a CLI selection command and usage cost estimates based on per-token pricing.
Suitable tasks: Switching to Hy4 Flash in the Command Code CLI or a session, and estimating usage costs at the rates shown on the page.
Unsuitable tasks: Treating Command Code's plans, prices, or model ID as access information for services such as Tencent Cloud TokenHub or OpenRouter.
Applicable model version: Tencent Hy4 Flash, model ID tencent/hy4-flash.
Applicable client, agent, or API: Command Code CLI; the page also says the model can be switched within a session.
Recommended reasoning tier and parameters: Not specified.
Run this in the Command Code CLI:
cmd --model tencent/hy4-flashYou can also enter /model in a session and select this model. The page says you can switch without interrupting the current context.
The FAQ says the model is available on the Go plan and above. The model page also lists Go ($1/month), GOAT ($10/month), Pro ($20/month), Max 10× ($100/month), and Max 20× ($200/month). The page does not specify the requirements for registration, login, or payment.
Confirm that the Command Code account qualifies for access on the Go plan or above.
At startup, run cmd --model tencent/hy4-flash, or run /model in an existing session and select the model.
The page lists rates of $0.07 per million tokens for input, $0.28 per million tokens for output, and $0.003 per million tokens for cached reads. Cost estimates should distinguish new input, cached reads, and output:
Estimated cost = new input tokens × 0.07/M + cached read tokens × 0.003/M + output tokens × 0.28/M
One Agent task example on the page uses 180K input tokens (75% cache hit rate) and 12K output tokens, for an estimated cost of $0.0069. This amount is calculated from the page's rates and assumptions; it is not a measured bill.
The model page shows the model ID as tencent/hy4-flash and the CLI command as cmd --model tencent/hy4-flash; the FAQ also lists /model as the in-session selection method.
The FAQ answers the plan requirement with “Available on Go and above.”
The page lists input at $0.07/M, output at $0.28/M, and cached reads at $0.003/M. It summarizes the effective input rate for an Agent loop as approximately $0.02/M. This effective rate depends on the cache share; the page says its estimate uses a median mix of about 30% new input and 70% cached reads.
The page's calculator defaults to 800 new input tokens, 180 output tokens, and 50K cached read tokens per request, and shows an estimated cost of about $0.0003 per request. For 30 days and $6 credits, it shows about 23K requests. The page does not explain how the $6 credits correspond to a plan, how credits are deducted, or what other assumptions the calculation uses, so this should be treated only as an example estimate from the page's calculator.
The page also lists cost estimates for quick queries, PR reviews, bug fixes, refactoring, and full-repository Agent runs. Apart from the Agent task card above, those table rows do not specify their cache ratios. Do not treat them as measured costs.
Both the Intelligence index and Coding index on the page are marked “not yet scored.” The page provides no quality benchmark scores for this model.
The command and plan eligibility information comes from the Command Code model page; it does not apply to other clients or providers.
Prices, cache ratios, and monthly request counts are estimates shown on the page. Actual costs depend on request token counts, cache hits, plan credit rules, and price changes. Cost estimates are not evidence of model quality or benchmark performance.
The page does not explain how to set the reasoning tier, temperature, or context limit in the client; configure an API key; open or pay for an account; or how the $6 credits correspond to a plan. This information is not specified.
This note records the page's configuration details and estimates. Command Code was not logged into, the CLI was not run, and no model request was sent. This does not establish account availability, command execution results, or model quality.
Open the model page to review the model ID, command, FAQ, and cost calculator. To verify actual account permissions and CLI behavior, run the command in an environment with an eligible Command Code plan and check the session model; this note did not perform that verification. Refer to the model page at the time of use for current prices and plan information.
Hy4 Flash