How much does Muse Spark Open cost?

Muse Spark Open is published as open weights (Muse Model License) — there is no per-token API price. You self-host the model and pay infrastructure cost (GPU rental and operating overhead). Effective cost on commodity 24-48GB GPUs typically lands at $0.10-$0.50 per million output tokens.

What is the context window of Muse Spark Open?

Muse Spark Open supports a context window of 1000K tokens (1,000,000 tokens) and can generate up to 66K output tokens per request.

What is Muse Spark Open best for?

Muse Spark Open is best for: Self-hosted multimodal agents (needs ~700GB GPU memory). It supports chat, vision, function_calling, code, reasoning capabilities. It is open-source under the Muse Model License license.

Muse Spark Open

Metaopen-source NewOpen Source

Meta's 15 Oct 2026 open-weights release, which Meta calls the most capable US open-weights model to date: 816B total / 42B active parameters, 62 layers, 256 experts with 8 active per token, 1M-token context, text, image, video and audio in, text out. Shares its base model with the closed Muse Spark 1.3 and adds safety post-training so it holds up as a standalone model without system guardrails. Weights (BF16 and NVFP4) are on Hugging Face behind a manual access request under the Muse Model License: free use, fine-tuning, distillation and redistribution, but a 'Built with Muse Spark Open' attribution, and any company with over $20M annual revenue that sells model access as an API must sign a separate agreement with Meta; running it needs at least 700GB of GPU memory at NVFP4. Artificial Analysis Intelligence Index v5.0.1 scores it 25 at xhigh effort at $2.02 per Index task, 90.6M output tokens and 172 tokens/s, with 13.28s to the first answer token. Meta publishes no first-party API price, so the row shows as self-host (third-party hosts list roughly $1.25/$4.25). Maximum output is not published; the figure shown is Muse Spark 1.3's.

Context Window

tokens

Max Output

66K

tokens

Input Price

Self-host

open weights

Output Price

Self-host

infra cost only

Speed

172

tokens/sec

Released

Oct 2026

2026-10-15

Blended Cost

—

self-host

Value Score

—

n/a self-host

Capabilities

ChatVisionFunction CallingCode GenerationReasoning

Benchmarks

Quality Index

Price vs Quality

Context Window in Context

▸ Muse Spark Open1M
GPT-4.11M
Gemini 2.5 Pro1M
Gemini 2.0 Flash1M
Llama 4 Maverick1M
GPT-5.51M
GPT-5.5 Pro1M
Claude Opus 4.71M

Bar length is logarithmic — each full step is 10× the tokens. Shown against the models with the nearest context windows, not the directory extremes.

Try in Playground Use via Swfte Connect Meta Docs

Compare With

Open Source: Licensed under Muse Model License

아직 게이트웨이에 없습니다

Muse Spark Open은(는) 공개 가중치 모델이라 직접 실행할 수 있습니다. Swfte Connect는 배포본을 이 디렉터리의 다른 모델과 동일한 API를 사용하는 관리형 엔드포인트로 만들어 줍니다.

요청

curl https://api.swfte.com/agents/v2/gateway/chat/completions \
  -H "Authorization: Bearer $SWFTE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "meta:muse-spark-open-816b-a42b",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Swfte Connect로 이 모델 배포하기

About Muse Spark Open

Muse Spark Open is a open-source AI model by Meta, released on October 15, 2026. It supports a context window of 1000K tokens and can generate up to 66K output tokens.

Muse Spark Open is published as open weights (Muse Model License) for self-hosting. There is no per-token API price. Cost depends on your inference infrastructure: GPU rental, throughput per GPU, and operating overhead. For commodity 24-48GB GPUs the effective cost typically lands in the $0.10-$0.50 per million output tokens range, well below the cheapest hosted alternatives.

Muse Spark Open is available as an open-source model under the Muse Model License license, meaning you can self-host it for predictable costs or use it through API providers like Swfte Connect.

Using Muse Spark Open with Swfte

Access Muse Spark Open through Swfte Connect, our unified LLM gateway. Connect gives you a single API for 50+ models, with automatic routing, cost optimization, and fallback handling. You can also try Muse Spark Open in our AI Playground before integrating.