Muse Spark Open
MetaReleased Oct 15, 2026Open weights · Muse Model Licensemuse-spark-open
- Intelligence (Epoch)
- 155 · 28th of 111
- Reasoning
- yes, with effort levels
- Knowledge cutoff
- January 2026
- Context
- 1M tokens
- Max output
- 256k tokens
- Input
- text, image, video, audio
- API features
- tools, structured outputs, JSON mode, seed, logprobs
Read Oct 23, 06:00 UTC from the hosts' published offers and OpenRouter, spec sheet from Meta · cutoff: dev.meta.ai
Intelligence for the price
9 models score higher for the same price or less; it ranks 10 of the 80 models at its price or below.
Epoch Capabilities Index against the lowest input + output price per M, log scale · prices Oct 23. Score: Epoch AI, CC BY 4.0, 155 with an interval of 153 to 158.
Against its rivals
The best of its price range and the top score
- Muse Spark Openthis model$5.50155
- Claude Opus 5.5top score$24.00169
- Gemini 3.7 Flashsame price range$4.50158
- Gemini 3.8 Flashsame price range$4.50157
Same level for 29× less: GLM-5.3-Flash at StreamLake, $0.19 per M input + output · their Epoch intervals overlap (153 against 155).
Bars: Epoch Capabilities Index (Epoch AI, CC BY 4.0), drawn from 100, below the lowest score in the catalogue · price: lowest input + output per M.
Speed by provider
Output speed of Muse Spark Open at each host, fastest first
- Modal263 tok/s0.22 s to first token
- Together AI155 tok/s0.39 s to first token
- BaseTen129 tok/s0.55 s to first token
- Fireworks AI76 tok/s0.44 s to first token
- Nebius46 tok/s1.50 s to first token
Speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · a snapshot of real traffic before the daily read, not a daily average · bars from zero.
Run it yourself
open weights, Muse Model License · too large for the 6 workstations in our guideSelf-host guide ›- Parameters
- 816B
- 42B active per token (MoE)
- Memory needed
- 490 GB
- in INT4 · 979 GB in INT8
- Smallest cloud setup
- 8× H100
- 640 GB · $17.20/h, about $12,556 a month
- Worth it above
- 2B tokens
- a month, against the cheapest API at $5.50 per M
| Quantization | Memory | Smallest cloud GPU | Quality |
|---|---|---|---|
| INT4 (Q4) | 490 GB | 8× H100 · $17.20/h | smallest footprint, minor quality loss |
| INT8 (Q8) | 979 GB | multi-node, over 640 GB | near-lossless |
| FP16 (full) | 1958 GB | multi-node, over 640 GB | reference quality |
At home: 490 GB in INT4 is beyond any desktop or Mac in our machines table, so it means a multi-GPU server.
Memory ≈ parameters × bytes per parameter × 1.2 · cloud price: on-demand rate, running 24/7 · builds: GGUF, AWQ, MLX, GPTQ, Ollama · break-even calculator
| Provider | Input | Output | Cache ↑ | tok/s | First token | Uptime 24h | Checked |
|---|---|---|---|---|---|---|---|
| Together AInvfp4 · 1M ctx | $1.25 | $4.25 | $0.15 | 155 | 0.39 s | 100% | Oct 23 |
| Nebiusfp4 · 1M ctx | $1.25 | $4.25 | $0.15 | 46 | 1.50 s | 94.1% | Oct 23 |
| Fireworks AI1M ctx | $1.25 | $4.25 | $0.15 | 76 | 0.44 s | 96.6% | Oct 23 |
| BaseTenfp4 · 1M ctx | $1.25 | $4.25 | $0.15 | 129 | 0.55 s | 100% | Oct 23 |
| Modal1M ctx · 256K out | $1.25 | $4.25 | $0.15 | 263 | 0.22 s | 100% | Oct 23 |
5 of 5 providers · prices per million tokens, read Oct 23, 06:00 UTC · speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · uptime: OpenRouter, last 24 h · under each host: weights precision, context and max output as served, the maker's value in brackets when the host's differs
| Provider | Weights | Context | Max output | Tools | Structured outputs | Reasoning |
|---|---|---|---|---|---|---|
| Together AI | nvfp4 | 1M | not stated by the host | yes | yes | yes |
| Nebius | fp4 | 1M | not stated by the host | yes | yes | yes |
| Fireworks AI | not checked yet | 1M | not checked yet | yes | yes | yes |
| BaseTen | fp4 | 1M | not checked yet | yes | no | yes |
| Modal | not checked yet | 1M | 256K | yes | yes | yes |
Weights, context and max output: read at each host's own catalogue or docs where its terms allow, otherwise as the host declares to OpenRouter; hover a cell for its source and date · tools, structured outputs, reasoning: as declared to OpenRouter · not verified by us
Muse Spark Open spec sheet, field by field, with sources
Muse Spark Open spec sheetmaker-declared
What Meta declares about the model itself — read from its own weights repository and documentation, not from any host. What each endpoint actually serves is in the pricing above: the two can differ, and both are true.
- Context window
- 1,048,576 tokens
- Parameters
- 816,000,000,000≈ 816.0B
- Active experts
- 8 of 256 active per tokenmaker's docs · read 2026-10-17
- Weight precision
- BF16maker's docs · read 2026-10-15
- License
- Muse Model License — open weights, conditionalcurated · read 2026-10-15Scope: the licence on the published weights, and nothing else. It is not the maker's acceptable-use policy, which is a separate document, and it is not the contract you sign with whichever host you call — hosts set their own terms, which this index does not read.
- Knowledge cutoff
- January 2026maker's docs · read 2026-10-15
- Modalities
- text · image · video · audio → textmaker's docs · read 2026-10-15
- Max output (official)
- not published by the makermaker's docs · read 2026-10-15
- Release date
- October 15, 2026maker's docs · read 2026-10-15
- Maker lifecycle status
- not published by the maker
meta-models/Muse-Spark-Open-816B-A42B · weights repository @ 0fb0247 · file not readable · read 2026-10-15 · 8 of 10 fields published by the maker
| Provider | Chatbot 1,000 in / 500 out | RAG / long context 8,000 in / 500 out | Batch summarize 4,000 in / 1,000 out |
|---|---|---|---|
| Together AI | $3.38 | $12.12 | $9.25 |
| Nebius | $3.38 | $12.12 | $9.25 |
| Fireworks AI | $3.38 | $12.12 | $9.25 |
| BaseTen | $3.38 | $12.12 | $9.25 |
| Modal | $3.38 | $12.12 | $9.25 |
What each host bills for 1,000 requests of a representative job
Your own workload
Other Meta models · all its models, hosts and history ›
- Llama 3.3 70B12 providers · from $0.42
- Llama 3.1 8B5 providers · from $0.06
- Llama 4 Maverick5 providers · from $1.00
- Llama 3.1 70B Instruct3 providers · from $0.80
- Llama 4 Scout3 providers · from $0.40
- Llama 3.2 3B Instruct2 providers · from $0.38