Inference providers · open weights · Meta
Muse Spark Open
Open-weights model published by Meta; the API prices below are set by the 3 hosts that serve it, not by Meta. There is no maker's list price to compare with: Meta publishes the weights, not a per-token rate. Each figure was read on that host's own pricing page and links to it, with the date it was checked.
What each host charges
3 hosts, every one of them at $1.25 per 1M input tokens and $4.25 per 1M output tokens: no spread at all, within the 1.15× that AI Signal treats as the same price. Pick a host on latency, region or tooling; the bill will be the same.
| Host | Input | Output | Cached input | Context served | Checked | Source |
|---|---|---|---|---|---|---|
| Together AI | $1.25 | $4.25 | $0.15 | 1.0M tokens | October 29, 2026 | together.ai |
| Fireworks AI | $1.25 | $4.25 | $0.15 | — | October 29, 2026 | fireworks.ai |
| Baseten | $1.25 | $4.25 | $0.15 | — | October 29, 2026 | baseten.co |
Mentioned in
2 posts on AI Signal cite Muse Spark Open. The sentence is the post's own, taken from where it links here.
-
AI Pricing This Week: OpenAI Shutdown, Mistral, Oct 19-25 · October 26, 2026
Fireworks' page for Muse Spark Open, read on October 26, no longer prints the limited-preview sentence that named October 29.
-
AI Pricing This Week: Gemini 4 Argon, Muse Wonder, Oct 12-18 · October 19, 2026
One open-weights model was added: Muse Spark Open, which Meta's model card describes as "an 816B total/42B active parameter open weights model with a context window of 1M tokens."
Recorded price changes
Every change AI Signal has recorded for a host's price of Muse Spark Open, with the date it took or takes effect and why. Changes of our own rule are listed too, and say so.
-
September 1, 2026 · Fireworks AI (all models on this host) · price up 1.5×
Since September 1, 2026 Fireworks adds a 50% surcharge to the base price of the Serverless models it marks as US-only. The Fireworks prices on AI Signal are the base prices, without the surcharge.
From Meta's model card
- Size: “Muse Spark Open is an 816B total/42B active parameter open weights model with a context window of 1M tokens.” (model card on Hugging Face, read October 17, 2026)
- What it is for: “Built for coding and agentic work, Muse Spark Open is natively multimodal.” (model card on Hugging Face, read October 17, 2026)
- License: “Licensee shall prominently display "Built with Muse Spark Open" in the user interface of each product or service that uses the Software or any Derivatives, or in the accompanying documentation, website, or settings menu.” (model card on Hugging Face, read October 17, 2026)
Specifications
| Weights published by | Meta |
|---|---|
| Context length (maker) | 1.0M tokens |
| Context served by hosts | 1.0M tokens |
| API model ids | Together AI: meta-models/Muse-Spark-Open-816B-A42B; Fireworks AI: muse-spark-open; Baseten: meta-models/muse-spark-open |
| Hosts with a verified price | 3 |
Host prices verified on October 29, 2026, each on the host's own pricing page linked in the table above, never from a reseller catalog or a third-party table. Maker and context length from Meta's model card, checked October 17, 2026 (Context length '1M at inference' as stated on the model card (read 2026-10-17), stored as 1048576 like the 128k of Llama 3.3 (131072); the repository is gated and its config.json could not be read without accepting the license. Release date on the card: 'October 15th, 2026'. Meta does not serve this model on its Model API: 'You don't call it over Model API.' (dev.meta.ai/docs/overview, read 2026-10-17)). How the data is checked.
Every open model, every host, one table →
When comparing hosts is worth it, and when it is not →