Muse Spark Open

The open-weight version of Meta's Muse Spark, an 816B mixture of experts (42B active) built for coding and agent work, with extra safety post-training, under the Muse Model License.

Params: 816B, 42B active
Context: 1M tokens
Released: Oct 2026
License: Muse Model License

Visit Meta Compare vs other LLMs →

Hosted API price

Cheapest via Together AI · per 1M tokens

USD

Input: $1.25 /1M
Output: $4.25 /1M
Hosted by: 2 providers

Hosted API pricing

How we price

Order

Providers are listed cheapest input rate first, then by output rate, a tie going to the creator's own listing. A provider that lists the model without a published rate sits at the end.

Prices

Base-tier, on-demand rates in USD per 1M tokens, converted at ECB reference rates where a provider publishes in another currency. Cached-input and batch rates show only where a provider publishes them. "Cheapest" marks the one provider with the lowest input and output pair; a tie wears no badge.

Cost

Input rate times the input tokens plus output rate times the output tokens, at the monthly volume set above the table. Cached-input, batch, long-context and reasoning-token billing are not modelled.

Transparency and funding

Affiliates: Affiliate links are marked. We may earn a commission if you click them, but commissions never affect the order.

Every provider serving Muse Spark Open, cheapest input first. Set your monthly volume to see what each would bill.

Input M Output M

Every provider serving Muse Spark Open, with its price per 1M input and output tokens and what the volume above would cost
Provider	Input /1M	Output /1M	Cached input /1M	Cost at 10M in + 2M out	Link
Together AI	$1.25	$4.25	$0.15	$21.00	Visit website
Meta Creator	Open weights, no hosted price				Visit website

Heads up: Base-tier, on-demand rates per 1M tokens; cached, batch and long-context tiers excluded. A provider may serve a shorter context or a quantized build than the creator's release. Verify before provisioning. More on how we price.

Estimated cost to self-host

Muse Spark Open needs about 453 GB of GPU memory at 4-bit with 32K context. The cheapest rental that fits is 8× A100 at about $10,771 a month.

That costs the same as roughly 6B tokens a month on Together AI's API. Self-hosting is more expensive below that volume.

Single node ≤ 8

What it costs to self-host Muse Spark Open at each precision: the memory it needs, the cheapest rental that holds it, that rental's month and the API volume it pays off at
Precision	Memory	Cheapest fit (32K context)(Full 1M context)	Cost /mo	Break-even vs API
4-bit INT4 / FP4	453 GB516 GB	8× A100	$10,771$10,771	6B tokens /mo6B tokens /mo
8-bit FP8 / INT8	819 GB851 GB	4× B300	$26,352$26,352	15B tokens /mo15B tokens /mo
16-bit FP16 / BF16	1,636 GB1,700 GB	8× B300	$52,704$52,704	30B tokens /mo30B tokens /mo

Estimates based on median on-demand rates for Nvidia GPUs. Memory is weights plus KV cache for one request, using FP8 KV cache where supported and FP16/BF16 otherwise. Break-even assumes a 5:1 input-to-output ratio. No guarantee of runtime support, usable performance, or that a matching quantized build exists. Pricing methodology.

Similarly priced models

The models nearest Muse Spark Open by blended rate, each at its own cheapest provider.

Models priced nearest Muse Spark Open per 1M tokens, each at its cheapest provider
Model	Blended /1M	Input /1M	Output /1M	Context	Cutoff	vs Muse Spark Open
GPT-6.1 Terra OpenAI	$1.6667	$1.00	$5.00	1M	Jun 2026	−5%
Qwen3.7-Max Alibaba Cloud	$1.6667	$1.25	$3.75	1M		−5%
Muse Spark 1.1 Meta	$1.75	$1.25	$4.25	1M		0%
Muse Spark 1.2 Meta	$1.75	$1.25	$4.25	1M		0%
Muse Spark 1.3 Meta	$1.75	$1.25	$4.25	1M		0%
Muse Spark Open This model Meta	$1.75	$1.25	$4.25	1M
GLM-5.3 Z.AI	$1.7583	$1.25	$4.30	1M		0%
Mistral Medium 3.5 Mistral	$2.50	$1.50	$7.50	256K		+43%
Qwen3.8-Max Alibaba Cloud	$2.5333	$1.90	$5.70	1M		+45%
Grok 4.5 xAI	$2.6667	$2.00	$6.00	500K	Feb 2026	+52%
Grok 4.6 xAI	$2.6667	$2.00	$6.00	500K	Jan 2026	+52%

USD per 1M tokens at each model's cheapest listed provider. Blended is the cost of 10M input plus 2M output tokens, spread over the 12M.

Runpod

Spin Up a GPU in Seconds - No Setup Required

Deploy AI Models Globally in 30+ Regions

Scale from 0 to 1000 GPUs on Demand

PyTorch, Tensorflow & Docker Ready

Learn more

Capabilities

What Muse Spark Open accepts and can do, as published by Meta.

Text Accepts and generates natural-language text.
Image Accepts images as input alongside text.
Video Accepts video as input.
Audio Accepts audio as input.
Reasoning Thinks before it answers, always on or as a switchable mode.
Function calling Connect to external tools, APIs, and systems.

Common questions

What is Muse Spark Open good for?

Coding agents and long multi-step agent work on infrastructure you run, with text, images, audio and video as input. Weights under the Muse Model License, on a multi-GPU node.

When is Muse Spark Open not a good fit?

Model-as-a-service businesses above the Muse Model License's revenue threshold, which need a separate agreement with Meta, and every product built on it must show a Built with Muse Spark Open notice. Self-hosting takes a multi-GPU node, and no maximum output or knowledge cutoff is published.

What is the cheapest way to run Muse Spark Open?

Hosted, unless you push serious volume. Together AI charges $1.25 in / $4.25 out per 1M tokens. The cheapest rental that fits is 8x A100 at $10,771 a month, which costs the same as about 6B tokens a month on that API.

Can I self-host Muse Spark Open?

Yes. The weights are released under the Muse Model License. At 4-bit it needs about 453 GB of GPU memory, which starts at roughly $10,771 a month on the cheapest rental that fits.