Meta: Muse Wonder

meta/muse-wonder · MetaTokenizer lineage measured

Tokenizer lineage (measured): Llama4
Entered catalogue: 12 October 2026
Last tested: 17 October 2026
Evidence confidence: Very high

Why very high: an exact fingerprint shared with 5 models and an identical API contract inside the same group: two signals of different kinds agree.

Evidence

✓ exact tokenizer signature shared with 5 models
✓ API-surface contract identical to a group member

Compare with Meta: Llama 4 Maverick

Report a correction

Identity statement

Meta: Muse Wonder is Meta's model. Its measured tokenizer sits with the Llama4 group, which is evidence that it builds on the Llama4 tokenizer. Tokenizer reuse is normal engineering, open vocabularies travel between labs, and this says nothing against Meta's authorship of the model itself.

What we checked

Four different kinds of evidence, and what each one showed

We checked	What we found	Where it came from
How it counts tokens	Counts every one of the 50 test strings exactly like 5 other models	tk_20b9d005
How its API is set up	Set up identically to 2 other listings, field for field	api_ed0a435a
How much it can read and write	Reads up to 1,048,576 tokens at onceOpenRouter's listing shows 943,718 output tokens, which is 90% of the context window, the figure it carries where the provider declares no maximum.	its own listing
Whether it thinks before answering	Always reasons, and cannot be turned off, at max, xhigh, high, medium, low, minimal, medium by default	its own listing
Who served our requests	Meta	we measured it, 17 October 2026

Shares this fingerprint

Exact signature first; template-boundary shifts of the same signature beneath

Meta: Muse Glimmer 30Bmeta/muse-glimmer-30bexact · 50/50 Meta: Muse Spark 1.1meta/muse-spark-1.1exact · 50/50 Meta: Muse Spark 1.2meta/muse-spark-1.2exact · 50/50 Meta: Muse Spark 1.3meta/muse-spark-1.3exact · 50/50 Meta: Muse Spark Openmeta/muse-spark-openexact · 50/50 Meta: Llama 4 Maverickmeta-llama/llama-4-mavericksame tokenizer · template shift Meta: Llama 4 Scoutmeta-llama/llama-4-scoutsame tokenizer · template shift

Closest measured models

Agreement across the mutually clean test strings

Meta: Muse Glimmer 30Bmeta/muse-glimmer-30b50/50 · 100%Meta: Muse Spark 1.2meta/muse-spark-1.250/50 · 100%Meta: Muse Spark 1.1meta/muse-spark-1.150/50 · 100%Meta: Muse Spark 1.3meta/muse-spark-1.350/50 · 100%Meta: Muse Spark Openmeta/muse-spark-open50/50 · 100%Meta: Llama 4 Scoutmeta-llama/llama-4-scout48/50 · 96%

Click a row to open the full comparison.

The measured fingerprint

What each test string cost this model, in tokens

We sent Meta: Muse Wonder fifty short pieces of text and recorded what each one cost it in tokens. The bars below are those costs. Two models built on the same tokenizer produce the same bars; a model built on a different one produces a different set, which is what makes this a fingerprint.

All 50 rows

English and whitespace

en-prose14

en-long17

spaces-203

spaces-603

tabs-203

newlines-204

mixed-ws6

Digits

digits-93

digits-124

digits-3010

digits-sep7

float-long9

CJK

zh-common11

zh-long12

zh-rare18

ja-kana9

ja-kanji12

ko10

Other scripts

ru10

ar11

he10

hi12

th8

el14

Emoji

emoji-basic10

emoji-skin20

emoji-zwj-family11

emoji-zwj-x333

emoji-flags16

emoji-prof19

Rare Unicode

math24

boxdraw21

combining10

cjk-ext-b16

surrogates37

zalgo31

rtl-mix6

Code

py-code25

py-indent15

json20

html16

regex42

camel6

snake6

Repetition and encodings

rare-word-x531

repeat-tok41

base6435

hex11

url17

uuid27

Overhead subtracted: 8 prompt tokens.

Declared record

What the catalogue claims about this model

Muse Wonder is a frontier reasoning model from Meta, designed for complex personal and professional agentic tasks. It accepts text, images, video, audio, and PDF documents and returns text, with...

Modalities: text+image+file+audio+video->text
Declared tokenizer: Other
Prompt price: $4.00 / M tokens
Completion price: $20.0 / M tokens

Supported parameters

include_reasoningmax_tokensreasoningreasoning_effortrepetition_penaltyresponse_formatstructured_outputstemperaturetool_choicetoolstop_ktop_p

Default settings

No values declared

Catalogue entry

History

Every observation, kept as taken

12 October 2026Enters the OpenRouter catalogue with no declared family.
17 October 2026Fingerprinted in the YFarmX catalogue sweep · 50 of 50 strings measured clean.
17 October 2026YFarmX assessment: consistent with the Llama4 family, very high confidence.