MetaReasoning modelMuse Spark 1.2
Muse Spark 1.2 is a Meta proprietary reasoning model, released 5 Aug 2026. Its best configuration (xhigh reasoning effort) ranks #76 of 751 on the BenchLeader Index at 62.7 ±4.4, in the upper half. The ± is the point: 154 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in reasoning (70) and lowest in coding (55). At $2.00 per million tokens blended it is pricier than most ranked models. Output speed of 242 tokens per second puts it in the fastest quarter, with a first answer in 22.4 s. It has been measured at 2 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 2 Oct 2026.
- Blended price
- $2.00/M
- $1.25 in · $4.25 out
- Output speed
- 242 tok/s
- measured by Artificial Analysis
- First answer
- 22 s
- first token 3.73 s
- Context
- 1.0M
- Full answer
- 24 s
- median, reasoning included
- Cost per run
- $0.975
- one full Intelligence Index run
- Released
- 5 Aug 2026
- Overall index63
- Reasoning70
- Coding55
- Agents & tools56
- Maths66
- Knowledge67
- Human preference69
- Multimodal66
- Long context65
- Composite67
Versions
Meta has shipped 4 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| Muse Spark 1.3max | 2 Sept 2026 | 67.8 | #21 |
| Muse Spark 1.2xhighthis page | 5 Aug 2026 | 62.7 | #76 |
| Muse Spark 1.1 | 9 Jul 2026 | 64.2 | #55 |
| Muse Spark | 8 Apr 2026 | 64.8 | #46 |
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| xhighbest | 62.7 | #76 | 242 tok/s | 22 s | $0.0018 | Agents & tools 56 · Coding 55 · Composite 67 · Human preference 69 · Knowledge 67 · Long context 65 · Maths 66 · Multimodal 66 · Reasoning 70 |
| not stated | – | – | 101 tok/s | 3.73 s | $0.0018 | Agents & tools 34 · Maths 52 · Reasoning 71 |
How its index has moved
8 Sept 2026 to 3 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| SimpleBench | – | 74.5%#9 | SimpleBench | |
| LMArena Hard Prompts | 1511#16 | – | LMArena | |
| LiveBench Reasoningnot in index | 90.0%#13 | – | LiveBench | |
| GPQA Diamond (AA)not in index | 90.4%#61 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 45.5%#46 | – | Artificial Analysis | |
| CritPt | 17.7%#65 | – | Artificial Analysis | |
| LMCAnot in index | 48.4%#61 | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | 94.7%#39 | – | Epoch AI Benchmarking Hub |
Coding
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| SciCode | 56.4%#28 | – | SciCode | |
| WeirdML | 60.3%#49 | – | WeirdML | |
| LMArena Coding | 1531#18 | – | LMArena | |
| LMArena WebDev | 1532#43 | – | LMArena | |
| LiveBench Codingnot in index | 77.5%#35 | – | LiveBench | |
| SciCode (AA)not in index | 57.4%#24 | – | Artificial Analysis | |
| IOI | 21.8%#34 | – | Vals AI | |
| SWE-bench (Vals)not in index | 86.6%#13 | – | Vals AI | |
| Code Migrationnot in index | 29.9%#35 | – | Vals AI | |
| Vibe Code Bench v1.1not in index | 79.1%#21 | – | Vals AI | |
| DeepSWEnot in index | 54.9%#38 | – | Epoch AI Benchmarking Hub | |
| FrontierSWEnot in index | 12.0%#15 | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | 7.1%#81 | – | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | 80.2%#44 | – | Artificial Analysis |
Agents & tools
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| APEX-Agents | – | 36.4%#34 | Mercor | |
| LMArena Agent | -3.3#37 | – | LMArena | |
| LiveBench Agentic Codingnot in index | 57.6%#19 | – | LiveBench | |
| Terminal-Bench 2.1 (Vals) | 69.7%#28 | – | Vals AI | |
| GDPval (AA)not in index | 49.1%#45 | – | Artificial Analysis | |
| τ²-Bench Banking (AA) | 34.9%#48 | – | Artificial Analysis | |
| CyberBenchnot in index | 69.5%#20 | – | Vals AI | |
| Finance Agent v2not in index | 60.6%#3 | – | Vals AI | |
| Harvey's Legal Agent Benchmarknot in index | 25.4%#1 | – | Vals AI | |
| Legal Research Benchnot in index | 43.8%#17 | – | Vals AI | |
| SkillsBenchnot in index | 53.0%#17 | – | Vals AI | |
| Tax Agent Benchnot in index | 56.9%#36 | – | Vals AI | |
| Terminal-Bench 4.0 (Vals)not in index | 6.1%#32 | – | Vals AI |
Maths
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| ProofBench | – | 43.0%#29 | Vals AI | |
| LiveBench Mathematicsnot in index | 91.2%#28 | – | LiveBench | |
| LMArena Maths | 1468#50 | – | LMArena |
Knowledge
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| SimpleQA Verified | 60.3%#15 | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | 76.5%#31 | – | LiveBench | |
| AA-Omniscience | 27.2#41 | – | Artificial Analysis | |
| MMLU-Pro | 88.3%#20 | – | Vals AI | |
| LegalBench | 85.3%#21 | – | Vals AI | |
| CorpFin | 70.9%#5 | – | Vals AI | |
| TaxEval | 80.4%#1 | – | Vals AI | |
| BioMysteryBenchnot in index | 64.8%#14 | – | Vals AI | |
| Excel Modeling Benchmarknot in index | 57.0%#37 | – | Vals AI | |
| MedCodenot in index | 49.4%#21 | – | Vals AI | |
| MedScribenot in index | 90.1%#5 | – | Vals AI | |
| Public Benefits Benchnot in index | 68.5%#8 | – | Vals AI | |
| GDP.pdfnot in index | 16.0%#31 | 12.0%#41 | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in index | 45.4%#88 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 66.7%#54 | – | Artificial Analysis |
Instruction following
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| LiveBench Languagenot in index | 78.6%#37 | – | LiveBench | |
| LiveBench Instruction Followingnot in index | 74.3%#11 | – | LiveBench | |
| LMArena Instruction Followingnot in index | 1477#24 | – | LMArena |
Human preference
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| LMArena Text | 1494#11 | – | LMArena | |
| LMArena Creative Writingnot in index | 1456#31 | – | LMArena | |
| LMArena Multi-turnnot in index | 1521#2 | – | LMArena | |
| LMArena Longer Queriesnot in index | 1494#22 | – | LMArena |
Multimodal
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| LMArena Vision | 1305#16 | – | LMArena | |
| MMMU-Pro (Vals)not in index | 86.1%#22 | – | Vals AI | |
| MortgageTaxnot in index | 65.4%#42 | – | Vals AI | |
| SAGEnot in index | 47.7%#30 | – | Vals AI |
Long context
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| AA-LCR | 79.0%#97 | – | Artificial Analysis |
Composite
| Benchmark | xhigh | not stated | Source | Trend |
|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | 155.0#29 | Epoch AI Benchmarking Hub | |
| LiveBench | 78.0%#19 | – | LiveBench | |
| AA Intelligence Index | 39.6#60 | – | Artificial Analysis | |
| Vals Indexnot in index | 49.3#27 | – | Vals AI | |
| Vals Multimodal Indexnot in index | – | 69.8%#6 | Vals AI |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Meta | 101 tok/s | 3.66 s | $1.25 | $4.25 | 1.0M | – |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.150 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0018 | $0.0014 | 23.6 s |
| Summarise a 30-page report | 12,000 / 600 | $0.018 | $0.0077 | 24.8 s |
| Code edit | 6,000 / 1,500 | $0.014 | $0.0089 | 28.5 s |
| Agentic coding session | 60,000 / 4,000 | $0.092 | $0.043 | 38.9 s |
| Structured extraction | 2,000 / 200 | $0.0034 | $0.0017 | 23.2 s |
See also
Data as of 4 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Muse Spark 1.2: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/muse-spark-1-2, data as of 4 Oct 2026.