Contributor tier lands on OpenRouter at $0.10/M input

Meta has made Muse Spark 1.2 available on OpenRouter through a contributor tier priced at $0.10 per million input tokens and $0.20 per million output tokens. The listing was announced by Meta on Threads, which frames the tier as an exchange: lower rates in return for permission to use traffic.

At a glance

• According to Meta, the contributor label sets the condition for the pricing: prompts and outputs passing through the tier may be used to improve Meta's products.
• The figure set is small: $0.10 per million input tokens, $0.20 per million output tokens, and a 1M context window that OpenRouter shows on the model's listing page.
• Meta claims the tier matches GPT-5.6 Terra performance while undercutting GPT-5.6 Luna on price, a comparison Meta draws against two named GPT-5.6 variants rather than against its own earlier models.

Pricing a model Meta compares to GPT-5.6 Terra at $0.10 per million input tokens reads as a distribution move rather than a margin one, and Meta appears to be buying reach and training traffic on a marketplace where developers sort models by cost per million tokens. For teams that already treat inference as a commodity, the calculus shifts from capability to compliance, since anything routed through the contributor tier carries Meta's stated right to reuse it.

Meta states the condition plainly: on the Contributor tier, prompts and outputs may be used to improve Meta's products. The post does not specify which products, how long data is retained, or whether the terms can be declined while keeping the listed rates.

The performance framing comes from Meta rather than from a published evaluation. In the Threads post Meta says the contributor tier delivers GPT-5.6 Terra performance at a price under GPT-5.6 Luna, without citing benchmark scores or the rates charged for either GPT-5.6 variant.

OpenRouter is a routing marketplace that exposes models from multiple vendors behind one API and lists them with per-token prices and context limits, so models from different vendors appear side by side with their rates and context figures on the same page.

OpenRouter lists Muse Spark 1.2 with a 1M context window, the same figure Meta gives for the model. Muse Spark 1.2 was officially released on August 5, 2026, and the OpenRouter listing does not show when the integration went live on the marketplace.

What the contributor terms leave open

No end date has been announced for the contributor pricing, and the post does not say whether Muse Spark 1.2 is offered at a different rate without the data terms, or how prompts and outputs are selected for reuse. The OpenRouter listing carries the rates and the 1M context window but no integration date, leaving the performance comparison as Meta's own claim.