|RSS

Announcements Chatter Podcasts Policy The week Models Connect your AI

Top stories

AI lab announcements, news reports and podcasts.

Announcements Chatter About Privacy Connect RSS

Top stories

|RSS

Announcements Chatter Podcasts Policy The week Models Connect your AI

Top stories

||Meta All models

Announcements Chatter Podcasts Policy The week Models Connect your AI

Muse Spark Open

Name: Muse Spark Open
Brand: Meta

Announcements Chatter Podcasts Policy The week Models Connect your AI

Meta modelsSpark Open Wonder Full Wonder Spark 1.3 Contributor Spark 1.3 Voice Transcribe Spark 1.2 Glimmer

So far

15 Oct
announced2026
1
post about itFirst-party
48
outlets reportedOn its posts
5
platforms carry itAMD, Amazon +3

So far

15 Oct
announced2026
1
post about itFirst-party
48
outlets reportedOn its posts
5
platforms carry itAMD, Amazon +3

Meta·All models

Meta models

Muse Wonder Full12 Oct

Top stories

Muse Spark Open · Meta — forck

||Meta All models

Announcements Chatter Podcasts Policy The week Models Connect your AI

Muse Spark Open

Announcements Chatter Podcasts Policy The week Models Connect your AI

Meta modelsSpark Open Wonder Full Wonder Spark 1.3 Contributor Spark 1.3 Voice Transcribe Spark 1.2 Glimmer

So far

15 Oct
announced2026
1
post about itFirst-party
48
outlets reportedOn its posts
5
platforms carry itAMD, Amazon +3

So far

15 Oct
announced2026
1
post about itFirst-party
48
outlets reportedOn its posts
5
platforms carry itAMD, Amazon +3

Meta·All models

Meta models

Muse Wonder Full12 Oct

Muse Wonder

12 Oct

Muse Spark 1.3 Contributor2 Sep

Muse Spark 1.32 Sep

+5 →

Launch

Meta· 10 days ago

Introducing Muse Spark Open, The Most Capable US Open‑Weight Model to Date

Meta releases open-weight model for desktop agentic work

Story of the week · 12–18 Oct Most coveredOpen-weight release

TestingCatalog: “ICYMI: Meta's Muse Spark Open is optimized for new Mac Studio”

Reporting · 48 outlets

Meta· 15 Oct

Introducing Muse Spark Open, The Most Capable US Open‑Weight Model to Date

Where to run it

AMD

Day 0 support is here for the first open weights model: Muse Spark Open.

Muse Spark Open 🤝 AMD Instinct GPUs. Day 0 support is here for the first open weights model: Muse Spark Open. Start building 👇 https://t.co/ePbUh6utxf

10 days ago

Muse Spark Open, an open-weight model for coding agents

Meta's 816B open-weight model optimized for coding agents

Baseten10 days ago

Meta Muse Spark Open is live on Together AI

Meta's open-weight coding model arrives on Together AI

Together AI10 days ago

Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta

Meta releases open-weight multimodal MoE model with 816B parameters

Hugging Face10 days ago

Meta’s Muse Spark Open now available in Amazon SageMaker JumpStart
Amazon10 days ago

Chatter

Artificial AnalysisReported launch
Meta's Muse Spark Open reaches 150M tokens/s per MW on NVIDIA GB300
Results from Muse Spark Open, Meta’s newly released 816B (42B active) open-weights model, are now live on AA-AgentPerf. In our initial testing, the model reaches up to ~150M tokens/s throughput per MW of power used on 4x NVIDIA GB300 or B300 GPUs
Muse Spark Open scores 39 on the Artificial Analysis Intelligence Index. It is an MoE with 62 layers, of which 46 use 2,048-token sliding-window attention, cutting the KV cache requirement at 1M context by ~75% vs. full attention. Meta published an NVFP4 checkpoint and AMD published an MXFP4 checkpoint, which we use for the NVIDIA and AMD hardware respectively
Initial results:
➤ NVIDIA B300 and GB300 lead, each reaching ~150M tokens/s per MW in throughput-oriented (lower per-user speed) configs. GB300 pulls ahead at higher output speeds (~108M vs ~92M at 100 tok/s), but both have only been tested in aggregated serving configs so far, and we expect GB300's lead over B300 to grow with multi-node disaggregated configs
➤ Serving configs are improving quickly. On the same B300 node, NVIDIA's tuned SGLang configs deliver ~1.6-1.8x the throughput per MW of the launch-week cookbook configs at 50-100 tok/s. The gains come from TP4 deployments and a prefill-decode interval tuned per concurrency, at the cost of longer time to first token
➤ AMD MI355X reaches 70M tokens/s per MW at 65 tok/s, ~0.9x B200 between 65 and 100 tok/s. AMD's Oct 22 recipe update (prefill-decode and scheduler intervals, more tokenizer workers, more KV cache for full-attention layers) lifted its line ~1.7x per MW over the earlier recipe
We are continuing to test Muse Spark Open serving configurations, including multi-node disaggregation, and will be updating the leaderboard each week with the best results and configs found so far. We will also be publishing deeper reviews of the progress made in optimizing kernels, inference engines and configs for Muse Spark Open

See all chatter →

About the model

Lab: Meta
Licence: Muse Model License with limited commercial terms
Context window: 1M-token context window
OpenRouter: Listed 15 October 2026meta/muse-spark-open

Ask your AI

Your brief, straight from the record

Connect once, then ask Claude or ChatGPT what changed for the tools you use. Every item links to its page on forck.live.

+ any MCP app

Connect your AI →

About the model

Lab: Meta
Licence: Muse Model License with limited commercial terms
Context window: 1M-token context window
OpenRouter: Listed 15 October 2026meta/muse-spark-open

Ask your AI

Your brief, straight from the record

Connect once, then ask Claude or ChatGPT what changed for the tools you use. Every item links to its page on forck.live.

+ any MCP app

Connect your AI →

So far

So far

Launch

Introducing Muse Spark Open, The Most Capable US Open‑Weight Model to Date

Day 0 support is here for the first open weights model: Muse Spark Open.

Muse Spark Open, an open-weight model for coding agents

Meta Muse Spark Open is live on Together AI

Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta

Meta's Muse Spark Open reaches 150M tokens/s per MW on NVIDIA GB300

About the model

About the model