Muse Spark 1.2 Lite Version Rollout: OpenCode Available for Free for a Limited Time

Dynamic Beating AI News Flash: Meta's Muse Spark 1.2 Contributor version is rapidly expanding to third-party platforms. Following the latest integration with OpenRouter, the pricing remains consistent with Meta's official pricing: $0.10 for input and $0.20 for output per million tokens, with a caching input cost of only $0.002. The regular Muse Spark 1.2, on the other hand, is priced at $1.25 for input, $4.25 for output, and $0.15 for caching.

The Contributor version is not a downgraded model. It utilizes the same model and capabilities, with the trade-off being that users' input and output can be used by Meta for training and product enhancement. In terms of output pricing, it is approximately 95% cheaper than the standard version.

There are already various applications leveraging it. OpenCode Zen has even offered the Contributor version as a limited-time free model. Command Code has also been integrated, allowing its use even with the basic Go package. Vercel AI Gateway and NanoGPT also provide APIs, with Vercel being able to seamlessly connect to tools such as Claude Code, Codex, Hermes, OpenCode, and OpenClaw.

Artificial Analysis has recently given Muse Spark 1.2 xhigh an Intelligence Index of 57. While the price is indeed remarkably low, the community's real-world feedback has been mixed. Some consider it a budget-friendly alternative to DeepSeek V4 Flash, while others have reported that long tasks and tool invocations are still not sufficiently reliable. It is suitable for open-source projects and low-cost experiments; however, for proprietary code, one must first consider whether they are willing to share data with Meta.