Meta Muse Spark and Llama models: 10M tokens, part open-weight
Meta
Services found: 6
| # | AI model | Capabilities | API $/1M (in / out) | Context window | Price / mo | Output $/1M | Overview | Visit site |
|---|---|---|---|---|---|---|---|---|
| 1 |  Muse Spark 1.2 Meta |    | $1.25 / $4.25 | 1.05M tokens | API only | 4.25 | View | |
| 2 | Muse Spark 1.1 Meta | | $1.25 / $4.25 | 1.05M tokens | API only | 4.25 | View | |
| 3 | Muse Spark 1.3 Meta | | $1.25 / $4.25 | 1.05M tokens | API only | 4.25 | View | |
| 4 | Muse Glimmer 30B Meta |  | $0.3 / $1.1 | 131.1K tokens | API only | 1.1 | View | |
| 5 | Llama 4 Maverick Meta |  | $0.2 / $0.696 | 1.05M tokens | API only | 0.696 | View | |
| 6 | Llama 4 Scout Meta | | $0.1 / $0.3 | 10M tokens | API only | 0.3 | View |
Browse by need
Meta retired the Llama brand during 2026. The paid API now sells Muse Spark at $1.25 input and $4.25 output per 1M tokens on a 1.05M window, with a Contributor tier at $0.10 and $0.20 for teams willing to let Meta train on their traffic.
Muse Glimmer 30B is the self-hosting option, published under Apache 2.0 and not served on any Meta endpoint. Older Llama 4 weights remain downloadable under the community licence but no longer appear in Meta own documentation.
Llama is goneThe brand disappeared from Meta docs; Muse replaced it.
Data-for-discount tierContributor pricing trades training rights for a tenfold cut.
Apache-licensed Glimmer30B weights for self-hosting, no hosted endpoint.
Models from this vendor
6 Meta models in this rating. Prices are input and output per 1M tokens from the vendor official price list; for dirhams multiply by 3.6725. Tap a card to open the model page.
Muse Spark 1.2
#1
Meta
in $1.25out $4.251.05M tokensText & chatVision
Muse Spark 1.1
#2
Meta
in $1.25out $4.251.05M tokensText & chatVision
Muse Spark 1.3
#3
Meta
in $1.25out $4.251.05M tokensText & chatVision
Muse Glimmer 30B
#4
Meta
in $0.3out $1.1131.1K tokensText & chatVision
Llama 4 Maverick
#5
Meta
in $0.2out $0.6961.05M tokensText & chatVision
Llama 4 Scout
#6
Meta
in $0.1out $0.310M tokensText & chatVision
How to choose
- Decide on the Contributor tier consciously — it is a tenfold discount in exchange for training rights.
- Use Glimmer 30B if you need self-hosting; it is the only Apache-licensed piece.
- Do not plan new work on Llama; it is absent from current Meta documentation.
- Compare Muse Spark against Terra and Sonnet on your own tasks before migrating.
Common mistakes
Building on Llama todayThe brand is retired from Meta docs even though old weights remain downloadable.
Taking Contributor pricing without reviewYour traffic trains Meta models on that tier.
Who this suits
Budget-constrained teamsContributor pricing at $0.10 input is among the cheapest 1M-context options.
Self-hostersGlimmer 30B under Apache 2.0.
Existing Llama usersThe migration path runs to Muse, not to a newer Llama.
Related filters
Text & chat228 models for text generation and chat
Vision156 models that read images
Agentic workflows5 models built for autonomous multi-step work
Long context 1M+82 models holding 1M tokens or more
Alibaba (Qwen)34 models in this rating
OpenAI30 models in this rating
Google DeepMind17 models in this rating
Frequently asked questions
What happened to Llama?
Meta replaced it with the Muse family in 2026. Llama 4 weights remain downloadable but are filed under history.
What is the Contributor tier?
Cheaper API access in exchange for Meta training on your traffic. The standard tier does not.
Can I self-host Muse?
Only Glimmer 30B, which ships under Apache 2.0.
Is Llama really discontinued?
It no longer appears in Meta model documentation, and llama.com redirects to the Meta developer site.
What does Contributor tier cost?
$0.10 input and $0.20 output per 1M tokens, versus $1.25 and $4.25 on standard.
Frequently searched
By capabilityReasoningCodingVisionLong context 1M+Free tierOpen weightsLow latencyRealtime voiceImage generation
By vendorAnthropic (13)OpenAI (30)Google DeepMind (17)xAI (8)Meta (6)DeepSeek (9)Mistral AI (11)Alibaba (Qwen) (34)Z.ai (12)Moonshot AI (5)
Rate this page
Help us improve