---
title: "Weekly unlimited LLM API plan for short projects | BroadRouter"
description: "When a one-week unlimited DeepSeek V4.1 Flash plan beats an hourly session, a monthly plan or pay-per-token, with the math for hackathons, sprints and migrations."
url: https://broadrouter.com/blog/weekly-llm-plan-short-projects/
lang: en
updated: 2026-11-14
publisher: BroadRouter (a BroadNet site); API by BroadNet Inference, https://inference.broadnet.ai
---

# A weekly LLM plan for short projects: Orbit vs hour vs month

By the BroadNet Inference team · 5 min read · Published 2026-11-14

Orbit is one week of unlimited DeepSeek V4.1 Flash for $15 per session, paid from your BroadNet balance, with no token meter and one request at a time. It fits a sprint, hackathon or migration. For an afternoon, Meteor at $1 an hour is cheaper; for continuous use, Full Moon at $60 beats back-to-back weeks ($64.29 over 30 days).

## Key takeaways

-   Orbit: $15 per session per week, unlimited DeepSeek V4.1 Flash, one request at a time.
-   Meteor: $1 per session per hour, starting at purchase. Full Moon: $60 per session per month.
-   Back-to-back Orbit weeks cost $64.29 over 30 days, so ongoing work belongs on Full Moon.
-   Orbit and Meteor are paid from your balance; Full Moon renews monthly on your card.
-   Buy one session per agent you run in parallel.

**On this page**

1.  What are the three unlimited plans?
2.  When is a week the right length?
3.  Which projects fit a week?
4.  How do you buy and use a week?
5.  Which plan fits which project?
6.  What should you know before you buy?
7.  Summary

Not every project lasts a month. A hackathon is a weekend. A codebase migration might be five intense days. A client prototype might be a one-week sprint. Paying per token for that kind of burst is unpredictable, and buying a monthly plan you will use for a week is wasteful. BroadNet Inference sells unlimited DeepSeek V4.1 Flash in three lengths so you can match the plan to the project. This post compares them and shows where the weekly plan, Orbit, fits.

## What are the three unlimited plans?

All three give the same thing: [DeepSeek V4.1 Flash](https://broadrouter.com/models/deepseek-v4.1-flash/) with no token meter, no monthly cap and no rolling window, a 1M-token context, and a typical speed of about 100 tokens per second. They differ only in length and how you pay.

| Plan | Price per session | Lasts | Paid by | Renewal |
| --- | --- | --- | --- | --- |
| Meteor | $1 | one hour, from purchase | account balance | buy another hour |
| Orbit | $15 | one week | account balance | buy another week before it ends, keep the same key |
| Full Moon | $60 | one month | card | renews monthly, cancel anytime |

With a founders code, Full Moon is $45 a month for the first three months.

A **session** runs one request at a time, the way a single agent works. If you run an orchestrator and three sub-agents in parallel, you need four sessions. [How sessions work](https://broadrouter.com/guides/unlimited-llm-api-sessions/).

## When is a week the right length?

### Orbit vs Meteor hours

Meteor is the cheapest way to try the plan on a real task: one hour, starting when you buy it. It is also the right choice for a single focused session, such as an evening of refactoring with a coding agent.

It stops being the right choice when the hours add up. Compare what you would spend on Meteor hours across the week with one Orbit week:

```text
Meteor hours in a week × $1   vs   $15
```

At today’s prices, the week costs the same as 15 Meteor hours. If your project needs more than about two working days of agent time in a week, or the agent needs to run unattended overnight, Orbit is cheaper and you don’t have to watch the clock.

### Orbit vs Full Moon

A month of back-to-back Orbit weeks works out to **$64.29** over 30 days, compared with **$60** for one Full Moon session. So:

-   **One to three weeks of work:** Orbit, bought week by week.
-   **Ongoing work, or you expect to keep going after the project:** Full Moon.
-   **Unsure:** start with an Orbit week. If you are still using it heavily at the end, move to Full Moon rather than buying week after week.

### Orbit vs pay-per-token

Per token, DeepSeek V4.1 Flash costs $0.175 per 1M input, $0.0055 per 1M cached input and $0.52 per 1M output. The rule is the same as for the monthly plan: if a week of your measured usage would cost more than $15 per token, the week wins.

For scale: our worked example of a full-time coding agent (1,000M input tokens a month at 80% cache hits, 50M output) comes to $65.40 a month per token. An agent working that hard for a single week is close to the Orbit price on its own, before counting reasoning-heavy days, retries or a second agent. A light chat assistant is far below it and should stay on pay-as-you-go. The full method is in [the break-even math](https://broadrouter.com/blog/flat-rate-vs-pay-per-token-llm-api/), and the [flat-rate vs per-token calculator](https://broadrouter.com/tools/flat-rate-vs-per-token-calculator/) runs it on your own token counts.

## Which projects fit a week?

### Hackathons and game jams

Two or three days of non-stop agent use is exactly what an unlimited week is for. Each team member uses their own account (plan keys cannot be shared), and each agent running in parallel gets its own session.

### Codebase migrations

Framework upgrades, language ports and large refactors often take one or two weeks of intense agent work and then stop. A coding agent re-reads large parts of the repository on every step, and the 1M-token context lets it hold a lot at once. On a per-token key that is a lot of input; on Orbit it costs the same whatever the agent reads. The [Aider guide](https://broadrouter.com/guides/aider-unlimited-deepseek/) shows a terminal agent pointed at the plan.

### Overnight and batch runs

Test generation, documentation passes and evaluation suites can run unattended for days. Orbit makes the cost of a long run known before it starts. For tips on keeping long runs healthy, see [overnight coding agent runs](https://broadrouter.com/blog/overnight-coding-agent-runs/).

### Client prototypes and sprints

When a project has a fixed budget and a fixed end date, a fixed price per week is easier to quote than a per-token estimate.

## How do you buy and use a week?

1.  Top up your balance. Orbit is paid from the balance; the free starting credit does not cover plans.
2.  Buy an Orbit session on the [BroadNet plans page](https://inference.broadnet.ai/plans?utm_source=broadrouter&utm_medium=referral&utm_campaign=blog-weekly-llm-plan-short-projects&utm_content=body) and create a plan key.
3.  Point your tool at `https://inference.broadnet.ai/v1` with model `broadnet/deepseek-v4.1-flash` and the plan key.
4.  Give each parallel agent its own session.
5.  Near the end of the week, decide: buy another week (same key), move to Full Moon, or stop.

Requests on a plan key never draw on your balance, and your pay-as-you-go keys keep working separately. If you also need another model during the week (vision on Gemma 4 VL 26B, for example), use a pay-as-you-go key for it.

## Which plan fits which project?

| Your project | Best fit |
| --- | --- |
| Try an agent on a real task for an evening | Meteor ($1); see [hourly keys and Meteor hours](https://broadrouter.com/blog/hourly-llm-api-key-explained/) |
| Weekend hackathon, one agent per person | Orbit ($15) per person |
| Two-week migration with two agents in parallel | Two Orbit sessions, renewed once |
| Ongoing daily development | Full Moon ($60) |
| Occasional chat, a few hundred requests a day | Pay-as-you-go |

## What should you know before you buy?

-   **One request at a time per session.** If your tool fires requests in parallel, one session won’t serve them at once, and BroadNet’s pages don’t say whether the extra requests wait or are refused. Configure the tool for one request at a time, or buy more sessions. [Running parallel agents on a flat rate](https://broadrouter.com/blog/run-parallel-coding-agents-flat-rate/).
-   **Reasoning still counts toward max\_tokens.** It costs nothing extra on the plan, but too low a limit still returns an empty answer.
-   **Fair use.** The plans are for your own agents and products. Sharing or reselling the key is not allowed, and use follows BroadNet’s [Acceptable Use Policy](https://www.broadnet.ai/aup).
-   **Speed is the real ceiling.** At about 100 tokens per second, one session generates a lot, but not infinitely. If a single agent is the bottleneck, add sessions rather than expecting more from one.

## Summary

Use **Meteor** for an hour, **Orbit** for a project measured in days, and **Full Moon** for work measured in months. Below the break-even, stay on pay-as-you-go. For the full comparison, see [what “unlimited” really means](https://broadrouter.com/blog/what-unlimited-llm-api-really-means/) and the [pricing page](https://broadrouter.com/pricing/).

Free tools for this

-   [LLM API cost calculator](https://broadrouter.com/tools/llm-api-cost-calculator/) Price a workload on every model at once, with cache hits, reasoning and a flat-plan check.
-   [Flat-rate vs per-token LLM cost calculator](https://broadrouter.com/tools/flat-rate-vs-per-token-calculator/) Break-even between an unlimited plan and paying per token, per agent.
-   [Prompt caching savings calculator](https://broadrouter.com/tools/prompt-caching-calculator/) Bill with and without caching, by prefix share and hit rate.
-   [Prompt token diff](https://broadrouter.com/tools/prompt-token-diff/) Two prompt versions side by side: changed lines, exact token delta and the monthly cost difference.

[All free tools](https://broadrouter.com/tools/)

## Terms in this article

[Token](https://broadrouter.com/glossary/token/)

A token is the unit of text a language model reads and writes: a word, part of a word, a punctuation mark or a byte sequence produced by the model's tokenizer. LLM APIs count, limit and usually bill usage in tokens.

[Pay-as-you-go](https://broadrouter.com/glossary/pay-as-you-go/)

Pay-as-you-go is a pricing model in which you pay only for what you use, metered per unit such as tokens, requests or audio minutes, with no subscription or minimum commitment. Most LLM APIs bill this way, from a prepaid balance or a periodic invoice.

[AI coding agent](https://broadrouter.com/glossary/coding-agent/)

An AI coding agent is a software tool in which a language model works on a codebase in a loop: it reads files, edits them, runs commands such as tests, reads the output and decides the next step until the task is done or a limit is reached.

[Prepaid credit](https://broadrouter.com/glossary/prepaid-credit/)

Prepaid credit is money added to an API account before use. Each request deducts its cost from the balance, and when the balance runs out, metered requests stop until you top up. It caps spend at what you have paid in.

[Cached input tokens](https://broadrouter.com/glossary/cached-input-tokens/)

Cached input tokens are prompt tokens that an LLM provider has recently processed and can reuse from a cache instead of computing again, usually a repeated prefix such as a system prompt or document. Providers bill them at a lower rate than new input tokens.

[Input and output tokens](https://broadrouter.com/glossary/input-and-output-tokens/)

Input tokens are the tokens a language model reads in a request, including the system prompt, conversation history, tool definitions and images. Output tokens are the tokens it generates in reply, including any reasoning. LLM APIs usually price them separately, and output usually costs more.

[max\_tokens](https://broadrouter.com/glossary/max-tokens/)

max\_tokens is an LLM API parameter that sets the most tokens a model may generate in one response. When generation reaches the limit, it stops and the response is marked as cut off, for example with finish\_reason set to "length".

[Subagent](https://broadrouter.com/glossary/subagent/)

A subagent is a separate agent that a main agent starts to handle one focused task. It runs with its own context window, system prompt and tool access, does the work, and returns only a result or summary to the main agent, which keeps the main conversation small.

## Questions

### Can I extend an Orbit week without changing my key?

Yes. BroadNet says to buy another week before the current one ends and you keep the same key.

### Can I pay for Orbit with the free starting credit?

No. The starting credit covers pay-as-you-go and hourly keys; plans are purchased separately. Orbit and Meteor are paid from a topped-up balance.

### Is Orbit a different model from Full Moon?

No. All three plans give the same model, DeepSeek V4.1 Flash, with unlimited tokens and a 1M-token context. They differ only in how long they last.

### How many sessions do I need for a team hackathon?

One per agent running at the same time. Each person uses their own account, because plan keys cannot be shared.

## Keep reading

-   [Flat-rate vs pay-per-token LLM API: the break-even math](https://broadrouter.com/blog/flat-rate-vs-pay-per-token-llm-api/)
-   [What "unlimited" really means in an LLM API plan](https://broadrouter.com/blog/what-unlimited-llm-api-really-means/)
-   [How to run parallel coding agents on a flat-rate LLM API](https://broadrouter.com/blog/run-parallel-coding-agents-flat-rate/)
-   [How to run coding agents unattended overnight, safely](https://broadrouter.com/blog/overnight-coding-agent-runs/)
-   [LLM API price changes in October 2026: 3 rises, no cuts](https://broadrouter.com/blog/llm-api-price-changes-october-2026/)
-   [Do LLM API prices only go down? 52 rises and how to budget for them](https://broadrouter.com/blog/llm-api-price-increases/)
-   [Guide: How the unlimited LLM API plans work for agents](https://broadrouter.com/guides/unlimited-llm-api-sessions/)
-   [Guide: Run Aider on unlimited DeepSeek V4.1 Flash](https://broadrouter.com/guides/aider-unlimited-deepseek/)

Try it on your real workload. New accounts start with $5 of credit for Gemma and transcription. DeepSeek, GLM and Qwen3.8 need a top-up.

[Get started](https://inference.broadnet.ai/account/?utm_source=broadrouter&utm_medium=referral&utm_campaign=blog-weekly-llm-plan-short-projects&utm_content=page-cta) ·

**Sources**

-   [BroadNet Inference unlimited plans and plan FAQ](https://inference.broadnet.ai/plans?utm_source=broadrouter&utm_medium=referral&utm_campaign=blog-weekly-llm-plan-short-projects&utm_content=sources)
-   [BroadNet Inference machine-readable plan prices (pricing.json)](https://inference.broadnet.ai/pricing.json?utm_source=broadrouter&utm_medium=referral&utm_campaign=blog-weekly-llm-plan-short-projects&utm_content=sources)
-   [BroadNet Inference API reference](https://inference.broadnet.ai/docs?utm_source=broadrouter&utm_medium=referral&utm_campaign=blog-weekly-llm-plan-short-projects&utm_content=sources)
