Command Code has introduced its GOAT plan, a $10-a-month coding subscription that includes $70 in usage credits and access to more than 30 open-weight and closed models. The company announced the plan in an August 5 post on X, positioning it as a larger step up from its $1 Go tier.
The company’s official pricing page currently lists GOAT at $10 per month plus a processing fee, with about 75,000 estimated requests and per-model allowances. That distinction matters: the $70 is not cash, a refund balance or unrestricted access to every model. It is the monthly usage value Command Code assigns across the plan’s model catalog.
Introducing the GOAT plan 🐐
Command Code GOAT plan is the best low-cost coding plan on the market today.
– $10/mo to get $70 credits. 7x what you pay
– 30+ open weight and closed models$1 Go got you started. GOAT gets you shipping.
Subscribe — be a GOAT.
by Command Code (@CommandCodeAI) August 5, 2026
A larger step up from Go
Command Code’s Go plan costs $1 per month and includes $10 in credits, according to the same pricing table. GOAT multiplies both the subscription price and the included usage, but the credit increase is the headline change for developers who found the entry tier useful for testing and too small for sustained work.
The company’s 7x calculation is straightforward: $70 of stated usage allowances divided by a $10 subscription. It should not be read as a guarantee that every subscriber will receive the same number of requests. Request volume depends on the model selected, the size of the prompts and the length of the coding session. Command Code’s own estimates put the typical GOAT mix at roughly 75,000 requests, but the number is an estimate rather than a fixed quota.
Credits are split by model
The current GOAT documentation lists 36 models on the plan. The catalog includes open models such as GLM, DeepSeek, Qwen, Kimi and MiniMax entries, alongside selected closed models with their own allowances. Command Code’s public pricing page also lists GPT-5.6 Luna, Grok 4.5 and Gemini 3.7 Flash among the models available under GOAT’s per-model rules.
The allowance is not identical across that list. The documentation currently shows up to $70 for GLM-5.2 and Tencent Hy3, $60 for DeepSeek V4 Flash and Kimi K2.7 Code, and $47 for MiniMax M3. Newer models can start with a smaller allowance while Command Code negotiates provider terms. The figures can change as those deals change, so the live plan documentation remains the useful reference for anyone comparing models such as DeepSeek V4 Pro.
Rolling limits temper the headline
GOAT also has shorter usage windows intended to stop a single burst from consuming the full monthly pool. Command Code documents a $14 limit over five hours, $35 over seven days and $70 over a billing month. Those limits are measured in usage value, not raw request counts, which is why cheaper models can support far more coding turns than expensive ones.
Subscribers can buy extra credits at model cost when they need more capacity. The documentation says those top-up credits roll over and do not expire, while the plan’s included credits reset with the billing cycle. That makes GOAT less like an unlimited flat-rate tool and more like a discounted bundle of model access with guardrails around heavy bursts.
Who GOAT is aimed at
The plan is aimed at solo developers who want one coding agent with a broad open-model menu and occasional access to premium models. Go remains the low-cost way to try the product, while Pro and Max offer larger or broader premium allowances. Command Code also lists a separate Provider API plan, so GOAT is primarily a coding-agent subscription rather than a general API package.
The announcement gives open-model users a clearer middle tier: enough included usage to build something meaningful, but with model-specific limits that keep the economics visible. Whether the 7x headline holds for an individual developer will depend on the models and workload they choose, not just the monthly price.
