# LunaRoute — compare/fireworks

Markdown twin for https://www.lunaroute.com/compare/fireworks/ — canonical HTML at https://www.lunaroute.com/compare/fireworks/.


<div class="pricing-guide-intro">

Fireworks AI is a high-performance inference platform with a catalog of hundreds of models. You can call supported serverless models per token, or deploy your own with a choice of GPU, precision and region, billed by the GPU-second. LunaRoute serves a curated set of open models tuned for agent work, for a fixed monthly price for included inference.

Choose Fireworks if you want control over how your models are deployed, fine-tuning, or enterprise compliance at scale. Choose LunaRoute if you run agents every day and want a predictable monthly bill without managing deployments.

</div>

<section class="pricing-guide-glance" id="at-a-glance">

## At a glance

Fireworks details are from its [pricing page](https://fireworks.ai/pricing), [model page](https://fireworks.ai/models/fireworks/glm-5p3) and [docs](https://docs.fireworks.ai/guides/ondemand-deployments), checked October 10, 2026. The [source list](#fireworks-sources) links the references for pricing, Fire Pass, deployment controls, tools and privacy.

<table class="pricing-guide-table" role="table">
<caption class="sr-only">LunaRoute and Fireworks AI at a glance</caption>
<thead role="rowgroup">
<tr role="row"><th scope="col" role="columnheader">What changes</th><th scope="col" role="columnheader">LunaRoute</th><th scope="col" role="columnheader">Fireworks AI</th></tr>
</thead>
<tbody role="rowgroup">
<tr role="row"><th scope="row" role="rowheader">How you pay</th><td role="cell" data-label="LunaRoute">Base coding plans: $99 to $499.99 per month, sized by priority concurrent requests</td><td role="cell" data-label="Fireworks AI">Per token on serverless; per GPU-second on dedicated deployments ($8 per GPU-hour for H100 or H200, before region premiums)</td></tr>
<tr role="row"><th scope="row" role="rowheader">GLM 5.3</th><td role="cell" data-label="LunaRoute">Included in coding plans. NVFP4 with vision, 512K context</td><td role="cell" data-label="Fireworks AI">Standard: $1.40 in / $4.40 out per million tokens ($0.26 cached), 1M context</td></tr>
<tr role="row"><th scope="row" role="rowheader">Deployment control</th><td role="cell" data-label="LunaRoute">We choose and tune the serving setup</td><td role="cell" data-label="Fireworks AI">You choose supported GPU type, count, precision and region combinations for the model</td></tr>
<tr role="row"><th scope="row" role="rowheader">Claude Code</th><td role="cell" data-label="LunaRoute">One command generates settings to apply; sessions are part of included inference</td><td role="cell" data-label="Fireworks AI">FireConnect or environment variables; standard hosted-model calls are billed per token, with Fire Pass and external routes separate</td></tr>
<tr role="row"><th scope="row" role="rowheader">Flat-rate coding option</th><td role="cell" data-label="LunaRoute">Coding plans, including production work and teams, subject to plan terms</td><td role="cell" data-label="Fireworks AI">Fire Pass: personal, non-production use, not open to new subscriptions</td></tr>
<tr role="row"><th scope="row" role="rowheader">Prompt retention</th><td role="cell" data-label="LunaRoute">Zero data retention by default for LunaRoute-hosted inference</td><td role="cell" data-label="Fireworks AI">Not stored for open models by default, with exceptions (see below)</td></tr>
<tr role="row"><th scope="row" role="rowheader">Fine-tuning</th><td role="cell" data-label="LunaRoute">No</td><td role="cell" data-label="Fireworks AI">Yes</td></tr>
<tr role="row"><th scope="row" role="rowheader">Tools around the models</th><td role="cell" data-label="LunaRoute">Web search, OCR, document conversion, image generation</td><td role="cell" data-label="Fireworks AI">Model inference, training and deployment, plus native WebSearch and WebFetch in supported coding tools</td></tr>
<tr role="row"><th scope="row" role="rowheader">Compliance</th><td role="cell" data-label="LunaRoute">Hosted-inference zero data retention by default; dedicated infrastructure on Enterprise</td><td role="cell" data-label="Fireworks AI">SOC 2 Type II and HIPAA support listed; Enterprise admins can enable ZDR enforcement</td></tr>
</tbody>
</table>
<p class="pricing-guide-table-note">The LunaRoute column describes hosted inference and base coding plans; other plans differ. Model and capacity limits still apply. Metered add-ons and external providers are separate. Zero data retention is a data-handling policy, not a compliance certification. Prices, features and availability can change.</p>

</section>

<div class="pricing-guide-reading">
<nav class="pricing-guide-contents" aria-label="In this comparison">
<p>In this comparison</p>
<ul>
<li><a href="#at-a-glance">At a glance</a></li>
<li><a href="#how-you-pay">How you pay</a></li>
<li><a href="#fire-pass">What about Fire Pass?</a></li>
<li><a href="#deployment-control">Deployment control</a></li>
<li><a href="#claude-code">Claude Code</a></li>
<li><a href="#data-retention">Data retention</a></li>
<li><a href="#choosing">Which fits your workload?</a></li>
<li><a href="#switching">Switching from Fireworks AI</a></li>
<li><a href="#fireworks-faq">FAQ</a></li>
<li><a href="#fireworks-sources">Sources</a></li>
</ul>
</nav>
<div class="pricing-guide-copy">
<section id="how-you-pay">

## How you pay

Fireworks' standard serverless pricing is per token, at the same list price as several other hosts for GLM 5.3: $1.40 per million input tokens, $4.40 per million output tokens, $0.26 for cached input. Priority and Fast modes cost more; see [serverless pricing](https://docs.fireworks.ai/serverless/pricing). For dedicated deployments you pay per GPU-second. The [pricing page](https://fireworks.ai/pricing) says there are “no extra charges for start-up times” and lists an H100 or H200 at $8 per GPU-hour. Region-restricted deployments carry a 1.5× premium; total cost also depends on GPU count, replicas and running time.

Both are fair models, and both scale with usage or provisioned GPU time. For an agent that resends context on every step and retries when it fails, that means the bill follows the agent. Cached-input rates can reduce that cost. A dedicated deployment avoids per-token billing but brings back the question of how busy you can keep the GPUs, since provisioned idle capacity still costs money ([Should you rent GPUs or use managed inference?](/guides/rent-gpus-vs-managed-inference/)).

LunaRoute's base coding plans are a fixed monthly price for 2 to 10 priority concurrent requests, with available overflow included. Long sessions and retries don't add to the included-inference bill, within model and plan limits. Overflow depends on available capacity; queue limits and timeouts still apply, and clients should use bounded retries. Metered add-ons and external providers are separate. More in [Concurrency vs token pricing](/guides/concurrency-vs-token-pricing/).

</section>

<section id="fire-pass">

## What about Fire Pass?

Fireworks has a flat-rate pricing option for coding too. Fire Pass gives “eligible users access to selected open-weight model routers for personal, non-production agentic coding”, with `kimi-fast-latest` currently resolving to Kimi K3 Fast as the default. As of October 10, 2026, Fireworks' docs say it “is experimental and is not currently open to new subscriptions” ([Fire Pass docs](https://docs.fireworks.ai/firepass)). Existing users and promo recipients can still manage their pass; covered routes, limits and terms apply.

LunaRoute's coding plans are flat-rate for included inference, including production work and teams, subject to plan and usage terms. You can sign up today.

</section>

<section id="deployment-control">

## Deployment control

This is where Fireworks is strongest. On a dedicated deployment you choose the accelerator (several NVIDIA and AMD options), how many, the precision, the region and how replicas scale, within the combinations supported for your model. Its [deployment docs](https://docs.fireworks.ai/guides/ondemand-deployments) describe validated deployment shapes; not every model runs on every GPU, and regional availability varies. That's real control over the tradeoffs between speed, cost and quality, and you can see what a deployment is made of.

See also https://www.lunaroute.com/llms.txt
