GOAT Plan

The Command Code GOAT plan is the best low-cost coding plan on the market today: unlimited coding on 30+ top open and closed models for $10/month. Your $10 buys $70 of credits - a 7x multiplier, the highest of any $10 coding plan. With deals, that stretches beyond $100: more than 10x what you pay.

See every plan side by side on Pricing & Limits, or pick a plan from Studio > Billing.

Why create the GOAT plan?

At Command Code, we're incessantly curious about 1) open models, 2) how to get the best value out of them, and 3) how to make that value accessible to everyone. The GOAT plan is our answer to all three.

While almost every other coding agent was built to serve closed models, we built Command Code to be the best coding agent harness for open models.

We started with the $1 Go plan. The sheer audacity of that plan was a hit, but it was also a bit of a tease: a great way to get started, not enough usage to last the month.

The GOAT plan fixes that - it's the best value in the coding market today, with enough credits to do something meaningful with open models.

Coding with open models comes with hard problems - cost, reliability, safety, and the sheer complexity of running them well. We set out to solve all of it, so you can pick an open model and just code.

From getting DeepSeek to beat Opus, to building infra with leading ~98% cache hit rates, to fixing AI slop with the built-in /design skill - Command Code has made its way to the top of the open model coding world, everywhere. The GOAT plan is the next step in that journey: open models, accessible to everyone.

Open models are already beating frontier closed models at coding. How rich is this moment if you think of it.

With the v1 release, which is a full rewrite of this 6-year-old codebase, it's now arguably one of the best, most mature coding agent harnesses. Check out the new Mods API - you can build anything you can imagine.

We can't wait to see what you build with Command Code.

We're also open sourcing Command Code later this month.

Let's connect on @CommandCodeAI and in our Discord community.

Models included

Every model below is included on the GOAT plan - switch between them any time with /model in interactive mode. Rates are per 1M tokens; the full catalog, including the premium models that need Pro or Max, lives on commandcode.ai/models.

Usage Limits

The GOAT plan includes the following usage limits:

• 5-hour limit - $14 of usage
• Weekly limit - $35 of usage
• Monthly limit - $70 of usage

How far each window goes depends on the model - cheaper models allow more requests, pricier models fewer. We have a usage estimation calculator. Estimated request counts per limit window:

Model	Requests / 5 hours	Requests / week	Requests / month
DeepSeek V4 Flash (latest)	39,000	97,400	195,000
GLM-5.2	947	2,370	4,740
GPT-5.6 Luna	10,400	25,900	51,800
Tencent Hy3	7,080	17,700	35,400
Qwen 3.7 Max	491	1,230	2,460
Qwen 3.7 Plus	3,020	7,540	15,100
Qwen 3.6 Plus	2,330	5,830	11,700
MiniMax M3	2,770	6,930	13,900
Kimi K2.7 Code	1,080	2,710	5,420
Qwen 3.8 Max	327	817	1,630
MiMo V2.5	19,500	48,700	97,400
DeepSeek V4 Pro	5,690	14,200	28,400
MiMo V2.5 Pro	5,700	14,200	28,500
Muse Spark 1.2	428	1,070	2,140
Muse Spark 1.2 Contributor	18,200	45,500	90,900
Kimi K3	196	490	980
Kimi K2.7 Code HighSpeed	181	452	904
Grok 4.5	144	360	719
GLM-5.2 Fast	138	346	691
Inkling	396	989	1,980
Inkling Small	709	1,770	3,550
Step 3.7 Flash	1,670	4,180	8,370
Step 3.5 Flash	3,510	8,770	17,500
Nemotron 3 Ultra	575	1,440	2,870

The GPT-5.6 Luna number is no typo: GOAT gives Luna the full $70 allowance - the biggest Luna allowance of any $10 coding plan - good for roughly 51,800 requests a month.

Model the math yourself with the calculator on Pricing & Limits - it knows every model's GOAT allowance.

The estimates assume a typical agent request - ~800 fresh input tokens, ~50,000 cache-read tokens, and ~125-200 output tokens depending on the model family - at the prices below per 1M tokens. Most of an agent's volume is cache reads, which is why cache hit rates matter so much. Browse every model on Available Models, with per-token rates on Pricing & Limits.

Every model on the GOAT plan is listed below:

Model	Input	Output	Cache Read	Cache Write	Monthly credits
GLM-5.2	$1.40	$4.40	$0.26	-	$70
GPT-5.6 Luna	$0.20	$1.20	$0.02	$0.25	$70
Tencent Hy3	$0.14	$0.58	$0.035	-	$70
Qwen 3.7 Max	$2.50	$7.50	$0.50	$3.13	$70
Qwen 3.7 Plus	$0.40	$1.60	$0.08	$0.50	$70
Qwen 3.6 Plus	$0.50	$3.00	$0.10	-	$70
DeepSeek V4 Flash (latest)	$0.14	$0.28	$0.0028	-	$60
Kimi K2.7 Code	$0.95	$4.00	$0.19	-	$60
MiniMax M3	$0.30	$1.20	$0.06	-	$47
MiMo V2.5	$0.14	$0.28	$0.0028	-	$30
Qwen 3.8 Max	$2.00	$6.00	$0.25	$2.50	$25
DeepSeek V4 Pro	$0.435	$0.87	$0.003625	-	$20
MiMo V2.5 Pro	$0.435	$0.87	$0.0036	-	$20

New models start at 2x credits i.e. $20 credits on the $10/mo GOAT plan until we negotiate better deals. That's still 2x your money at public API prices, and the allowance updates automatically the moment a deal starts - no codes, no toggles.

Model	Input	Output	Cache Read	Cache Write	Monthly credits
Muse Spark 1.2	$1.25	$4.25	$0.15	-	$20
Muse Spark 1.2 Contributor	$0.10	$0.20	$0.002	-	$20
Kimi K3	$3.00	$15.00	$0.30	-	$20
Kimi K2.7 Code HighSpeed	$1.90	$8.00	$0.38	-	$20
Grok 4.5	$2.00	$6.00	$0.50	-	$20
GLM-5.2 Fast	$3.00	$10.25	$0.50	-	$20
Inkling	$1.00	$4.05	$0.17	-	$20
Inkling Small	$0.50	$1.20	$0.10	-	$20
Step 3.7 Flash	$0.20	$1.15	$0.04	-	$20
Step 3.5 Flash	$0.10	$0.30	$0.02	-	$20
Nemotron 3 Ultra	$0.60	$2.40	$0.12	-	$20

Speed variants are separate models with their own credits: GLM-5.2 Fast and Kimi K2.7 Code HighSpeed carry the standard 2x credits ($20 on the $10 GOAT plan), not their base model's boosted allowance.

We're constantly negotiating better terms with providers, so the allowances above can change at any time. When a deal starts, your allowance updates by itself.

How is 7x possible?

You pay $10 and code with up to $70 of per-model credit allowances. We benchmarked the GOAT plan to be the best value on the market. This is possible through phenomenal harness and inference engineering. We work directly with model teams and providers to negotiate capacity for the models we recommend, and we run our infrastructure at leading ~95-98% cache hit rates.

Most of a coding agent's volume is cache reads, so serving the cache well makes every request dramatically cheaper - and those savings come back to you as bigger credits.

Deals with the best pricing apply automatically. Many providers and coding agents charge up to 400% over model list prices, while we regularly offer deals matching - and sometimes beating - the labs' own API prices. We run our own infrastructure and have deep experience running open models at scale, which lets us offer better pricing than most providers.

That's also why credits vary by model. Where we've negotiated capacity, credits are boosted - up to the full $70 on models like GLM-5.2, GPT-5.6 Luna, Tencent Hy3, and the Qwen 3.6/3.7 family. New models, and models whose public pricing already discounted via deals, carry the 2x credits ($20 credits on the $10/mo GOAT plan) and move up as we land better terms. We regularly move capacity to the latest models that perform better.

What you get

• Every major open model: Switch any time with /model in interactive mode. Browse them all on Available Models, with per-token rates on Pricing & Limits.
• Usage limits that scale: 20% of your monthly usage in any 5 hours, 50% in any 7 days - see Usage Limits.
• Extra credits any time: Buy pay-as-you-go credits at model cost - they unlock every model (including premium), roll over, and never expire.
• Global availability: Open-source models run on infrastructure in the US, EU, and Singapore for reliable access worldwide.
• Zero data retention available: Most models are ZDR by default - agreements are renewed monthly and can take time for brand-new models. You can also enforce ZDR on every request (which can change model prices based on the provider, as explained in Pricing & Limits).

Usage beyond plan limits

Your budget refreshes at the start of every billing cycle, with every model discount already baked into the allowances - no codes, no toggles.

Buy extra credits at model cost any time: they roll over, never expire, and bill at the model's regular rate, since the bigger allowances apply to your monthly budget only.

Past a limit, requests fall back to those credits - and without them, paid models pause until the window or cycle resets while the free models keep working.

Get started

1. Install: npm i -g command-code
2. Sign in and start coding with cmd (macOS/Linux) or cmdc (Windows)
3. Pick the GOAT plan from Studio > Billing or the pricing page

Need premium models? Compare the Pro and Max plans, or use the pay-as-you-go Provider API. Full comparison and FAQs live on Pricing & Limits.

FAQs