Muse Spark Open vs RouterPlus

Muse Spark Open

Muse Spark Open runs on the owner's hardware in seven runtimes, with OpenAI-compatible servers in vLLM and SGLang, standard tool calls, a 1M-token context and Meta's own serving docs. Serving the 537 GB NVFP4 checkpoint takes four B300 GPUs, Meta's download is gated by manual review, and the licence adds attribution and a revenue threshold for model resellers.

RouterPlus

The docs are written for agents, every model price is public, and each response carries its own cost. No terms of service, privacy policy, status page or changelog were found. Signup adds no free credit, and the chat catalogue is Claude and GPT only.

Muse Spark Open

1. Start vLLM or SGLang with --reasoning-parser muse --tool-call-parser muse (vLLM also --enable-auto-tool-choice), or tool calls come back as raw text
2. Set reasoning_effort to low or medium for short tasks. The default is xhigh, reasoning can't be turned off, and other values return an error
3. Stop only on <|eot|> and end of text. Stopping on <|eom|> cuts off tool calls and final answers
4. Stream long generations and leave max_tokens room for reasoning, or content can come back empty with finish_reason length
5. Don't call Meta Model API for this model. Use a self-hosted server or a hosted provider's model string, such as meta-models/muse-spark-open

RouterPlus

1. Send max_tokens on every request. Without it the balance hold assumes up to 32,768 output tokens and can return insufficient_quota on a small balance
2. Switch on error_type. Retry rate_limit after retry-after, and stop on insufficient_quota, which needs a person to add credit or raise a cap
3. Don't add a retry loop on top of the SDK's. There is no idempotency key, so a retry that reaches a provider is billed again
4. Copy model ids from GET /v1/models or /api/models.json. An unknown id is a 404 and the gateway never substitutes one
5. Use /v1/messages for Claude when you need thinking or server tools, and /v1/chat/completions with response_format only for GPT ids