Name: Muse Spark Open
Price: 1.25 USD
Author: Meta

Rank	Model	Overall score	Artificial Analysis	LiveBench	Epoch AI	ARC-AGI	Scale SEAL	Vals AI	Terminal-Bench	EQ-Bench	SimpleBench	BenchLM
1	Claude Opus 5.5 Anthropic	92.4	#1	#2	#1	#3	—	#2	#1	—	#1	#1
2	Gemini 4 Argon Google	91.1	#6	—	#5	#5	—	#1	—	—	—	#7
3	Muse Wonder Full Meta	91.1	#3	—	—	—	—	#5	—	—	—	—
4	GPT-6 Astra OpenAI	90.9	#5	#4	#2	#1	#1	#7	#3	—	#4	#3
5	Claude Fable 5.1 Anthropic	90.7	#4	#1	#4	#7	#2	#4	#5	—	#2	#4
6	Claude Sonnet 5.5 Anthropic	90.4	#2	#21	#6	—	—	#2	#2	—	#14	#5
7	GPT-6.1 Sol OpenAI	90.2	#8	#5	#3	#2	—	#9	#3	—	#5	#2
8	Claude Opus 5 Anthropic	89.5	#9	#11	#7	#6	—	#6	#6	#1	#9	#8
9	Claude Fable 5 Anthropic	89.2	#10	#3	#9	#9	—	#8	#8	#2	—	#6
10	Muse Wonder Meta	88.7	#7	—	—	—	—	#10	—	—	—	—
11	GPT-6 Sol OpenAI	86.7	#12	#13	#8	#8	—	#13	#7	—	#17	#9
12	GPT-5.6 Sol OpenAI	85.8	#13	#8	#11	#4	—	#12	#11	#9	#25	#10
13	GPT-5.5 Pro OpenAI	85.5	—	—	#10	#12	—	—	—	—	#11	—
14	Muse Spark 1.3 Meta	85.2	#11	#5	#21	—	—	#11	#20	—	#8	—
15	GPT-5.4 Pro OpenAI	84.1	—	—	#13	#16	—	—	—	—	—	—
16	MiMo V2.6 Pro Xiaomi	83.5	#15	—	—	—	—	#14	—	—	—	—
17	GPT-5.5 OpenAI	82.6	#33	#10	#15	#11	—	—	—	#4	#21	#16
18	Gemini 3.8 Flash Google	81.6	#26	#33	#19	#9	#5	#16	#17	—	#6	#17
19	Grok 4.7 xAI	81.5	#14	#22	#39	#27	—	#16	#10	—	—	—
20	GLM-5.3 Z.ai (Zhipu)	80.5	#17	#31	#30	—	—	#19	#9	—	#24	#18
21	Claude Opus 4.8 Anthropic	80.0	#23	#29	#16	#21	—	#15	#14	#6	#25	#13
22	DeepSeek V4.1 Flash DeepSeek	79.9	#29	#8	#31	#20	—	#25	—	—	#23	#25
23	GPT-5.4 OpenAI	79.8	—	#17	#23	#19	—	—	—	#7	—	#21
24	Kimi K3 Moonshot AI	79.7	#20	#13	#17	#31	—	#29	—	#3	#35	#14
25	GPT-5.6 Terra OpenAI	79.6	#22	#20	#12	#15	—	#21	#15	#11	#55	#11
26	Grok 4.6 xAI	79.5	#18	#17	#27	#24	—	#22	#16	—	#14	#19
27	Qwen3.8 Max Alibaba (Qwen)	79.2	#16	#16	#34	—	—	#34	#13	—	—	#15
28	Gemini 3.7 Flash Google	79.1	#30	#15	#18	#12	—	#25	#23	—	—	#20
29	MiMo V2.6 Flash Xiaomi	78.6	#37	—	—	—	—	#20	—	—	—	—
30	Step 5 StepFun	78.1	#19	—	—	—	—	#31	—	—	—	—
31	Claude Opus 4.7 Anthropic	78.1	—	#28	#26	#18	#10	—	—	#5	#28	—
32	Muse Spark 1.2 Meta	77.6	#28	#17	#35	—	—	#32	—	—	#16	#24
33	Claude Haiku 5.5 Anthropic	77.6	#21	#48	#14	#29	—	#18	—	—	—	#12
34	Qwen 3.8 Flash Next Alibaba (Qwen)	76.3	#27	#29	—	—	—	—	—	—	—	—
35	Claude Sonnet 5 Anthropic	75.0	#35	#32	#24	—	—	#23	#21	#10	#36	#26
36	Grok 4.20 xAI	73.9	—	—	#49	#26	—	—	—	—	—	—
37	GPT-6 Luna OpenAI	73.5	#36	#49	#27	#36	—	#27	#19	—	—	#22
38	Muse Spark Open Meta	73.3	#31	—	—	—	—	#37	—	—	—	—
39	GLM-5.3 Flash Z.ai (Zhipu)	73.1	#23	#52	#45	#25	—	#40	#12	—	—	—
40	Claude Opus 4.6 Anthropic	72.9	—	#39	#33	#23	#11	—	—	#13	#22	#31
41	DeepSeek V4 Pro DeepSeek	72.7	#39	#22	#32	#29	—	#36	—	#18	#53	—
42	Muse Spark Meta	72.5	—	—	#48	—	#7	—	—	—	—	#37
43	Muse Spark 1.1 Meta	72.4	#44	#35	#41	—	—	—	—	#8	—	#23
44	DeepSeek V4 Flash DeepSeek	72.0	#41	#40	#37	#27	—	#35	—	—	#33	—
45	GPT-5.6 Luna OpenAI	71.9	#38	#42	#25	#35	—	#24	#18	#19	#57	#27
46	Grok 4.5 xAI	71.7	#32	#33	#42	#39	—	#39	#21	—	#20	#28
47	Gemini 3.1 Pro Google	71.3	#51	#25	#36	#17	#3	#43	—	#21	#10	—
48	Gemini 3.5 Flash Google	70.6	#47	#37	#38	#21	—	#38	—	#24	#12	#32
49	Mistral Large 4 Mistral AI	70.1	#33	#51	#70	—	—	#33	—	—	—	#33
50	Gemini 3.6 Flash Google	69.6	#42	#42	#40	#31	—	—	—	—	—	#29

Rank	Provider	Overall score	Top model	Models
1	Anthropic	86.8	Claude Opus 5.5	31
2	OpenAI	84.5	GPT-6 Astra	54
3	Google	80.2	Gemini 4 Argon	35
4	Meta	76.1	Muse Wonder Full	27
5	xAI	69.0	Grok 4.7	13
6	Moonshot AI	66.8	Kimi K3	5
7	DeepSeek	66.4	DeepSeek V4.1 Flash	22
8	Alibaba (Qwen)	66.1	Qwen3.8 Max	46
9	Z.ai (Zhipu)	64.2	GLM-5.3	9
10	StepFun	61.7	Step 5	1
11	Xiaomi	61.3	MiMo V2.6 Pro	4
12	Mistral AI	40.9	Mistral Large 4	27
13	Tencent	40.0	—	2
14	Thinking Machines	38.1	Inkling	2
15	MiniMax	34.9	MiniMax M3	5
16	NVIDIA	30.4	Nemotron 3 Ultra 550B A55B	12
17	Amazon	18.7	—	3
18	Microsoft	12.8	Phi-4	6
19	Cohere	11.3	Command R	3
20	Inception	8.8	Mercury 2.5	2

A provider’s result on each site is that of its best model there.

Performance vs price

Top left = cheap and strong. Click a point for details.

Best value

Overall score ÷ price

Top open-weight models

Models with public weights

Fastest models

Output tokens per second

By country

Top 50 models

Agreement between leaderboards

Rank correlation on shared models (1 = identical).

How the ranking works

Collect — Official data from the 11 leaderboards below is fetched every hour. Scores marked as estimates are not used.
Match — Different spellings of the same model are merged. For variants such as reasoning effort, each site’s best result is used.
Standardize — Each site’s main metric is standardized within that site, keeping score gaps rather than just ranks.
Equate — A factor model, bridged by models listed on several sites, corrects for each site’s difficulty and model mix. Noisier sites get less weight.
No penalty for missing data — Models are scored only from the sites that rated them. Uncertainty is shown as a 90% interval. Models rated on 2+ sites are ranked.
Arena — Arena (formerly LMArena) is used for domain rankings only, not for the overall score. New models are added automatically.

Leaderboard	Rank	Published value
Artificial Analysis	#31 / 141	38.9
LiveBench	Not rated	—
Epoch AI	Not rated	—
ARC Prize (ARC-AGI)	Not rated	—
Scale AI SEAL	Not rated	—
Vals AI	#37 / 47	46.6
Terminal-Bench	Not rated	—
EQ-Bench	Not rated	—
SimpleBench	Not rated	—
BenchLM.ai	Not rated	—

Muse Spark Open: benchmarks and ranking

Rank on each leaderboard

Domain scores