Muse Spark 1.2 (xhigh): benchmark scores, price and ranking
Muse Spark 1.2 (xhigh) from Meta AI ranks #113 of 743 LLMs on the LLMs Tiger aggregate benchmark score (51.5), based on 13 benchmark results captured up to 2026-10-13.
- Organisation
- Meta AI
- Type
- Frontier model
- Open weights
- No
- Released
- 2026-08-05 (source: epoch)
- Cheapest API price
- $1.25 in / $4.25 out per million tokens (via meta, observed 2026-10-13T19:47Z)
- Aggregate rank
- #113 of 743 (score 51.5)
Relative to other models, Muse Spark 1.2 (xhigh) is strongest on DTBench (#11 of 146) and weakest on FrontierSWE (#18 of 19).
Across the 13 benchmarks where at least ten models are scored, Muse Spark 1.2 (xhigh) is in the top 10% on 1 and in the top half on 9.
Its cheapest listed input price, $1.25 per million tokens, is lower than that of 26% of the 427 models with a listed price.
Muse Spark 1.2 (xhigh) benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| DTBench | 94.7% | #11 of 146 | Epoch AI Benchmarking Hub | 2026-10-13 |
| LMArena Vision Elo | 1,293 | #3 of 22 | LMArena | 2026-10-13 |
| LMArena Text Elo | 1,494 | #4 of 28 | LMArena | 2026-10-13 |
| LMCA | 48.4% | #17 of 108 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SimpleQA Verified | 60.3% | #11 of 69 | Epoch AI Benchmarking Hub | 2026-10-13 |
| WeirdML | 60.3% | #19 of 109 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SciCode | 56.4% | #19 of 106 | Epoch AI Benchmarking Hub | 2026-10-13 |
| CritPt | 17.7% | #25 of 115 | Epoch AI Benchmarking Hub | 2026-10-13 |
| WebDev Arena | 1,534 | #18 of 67 | Epoch AI Benchmarking Hub | 2026-10-13 |
| DeepSWE | 54.9% | #15 of 19 | Epoch AI Benchmarking Hub | 2026-10-13 |
| GDP.pdf | 16.0% | #20 of 25 | Epoch AI Benchmarking Hub | 2026-10-13 |
| LMArena Code Elo | 1,532 | #25 of 30 | LMArena | 2026-10-06 |
| FrontierSWE | 12.0% | #18 of 19 | Epoch AI Benchmarking Hub | 2026-10-13 |
About Muse Spark 1.2 (xhigh)
Muse Spark 1.2 (xhigh) is a frontier model from Meta AI. It was released on 2026-08-05, according to Epoch AI. The cheapest API listing we track charges $1.25 per million input tokens and $4.25 per million output tokens, through meta. We hold 13 benchmark results for it across 4 areas, and 1 is in the top tenth of its board.
Where Muse Spark 1.2 (xhigh) stands
On our aggregate score, Muse Spark 1.2 (xhigh) is #113 of 743 ranked models, in the top quarter. Its score is 51.5; the scale is centred on 50, so it sits 1.5 points above the typical model. That beats 85% of the ranked field. Among the 28 ranked Meta AI models it is #4. The next one up is Muse Spark 1.1 (xhigh) at 52.4. The next one down is Llama 3.1-405B at 51.2.
Muse Spark 1.2 (xhigh), benchmark by benchmark
On DTBench, Muse Spark 1.2 (xhigh) scores 94.7% and ranks #11 of 146. Only 10 models score higher, led by Claude Opus 5.5 (max) at 98.9%, which is 4.3 points ahead. It covers dynamic tool-use benchmark. The board median is 68.8%, so Muse Spark 1.2 (xhigh) is 25.9 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
On LMArena Vision Elo, Muse Spark 1.2 (xhigh) reaches 1,293. It ranks #3 of 22, ahead of 86% of the field, and 3 points behind Gemini 3.7 Flash (high) at 1,296. The results either side are Muse Spark (1,294) and GPT-6.1 Sol (max) (1,291). It covers crowd-voted image understanding. The board median is 1,279, so Muse Spark 1.2 (xhigh) is 14 points above it. Source: LMArena, captured 2026-10-13.
Muse Spark 1.2 (xhigh) manages 1,494 on LMArena Text Elo, good for #4 of 28. gemini-4-argon-high leads at 1,525, 31 points further up, and 86% of models sit below. The results either side are Gemini 3.8 Flash (high) (1,495) and Muse Spark 1.3 (max) (1,494). It covers head-to-head chat preference (crowd votes). The board median is 1,478, so Muse Spark 1.2 (xhigh) is 16 points above it. Source: LMArena, captured 2026-10-13.
LMCA puts Muse Spark 1.2 (xhigh) at 48.4%, #17 of 108. That beats 84% of the models tested, but 19.9 points short of Claude Opus 5.5 (max) (68.2%). The results either side are Grok 4.6 (xhigh) (48.5%) and Claude Sonnet 4.6 (max) (46.5%). It covers language-model capability assessment. The board median is 29.2%, so Muse Spark 1.2 (xhigh) is 19.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
On SimpleQA Verified, Muse Spark 1.2 (xhigh) reaches 60.3%. It ranks #11 of 69, ahead of 84% of the field, and 15.7 points behind gemini-4-argon-high at 76.0%. The results either side are GPT-6 Sol (max) (60.7%) and Claude Opus 5 (max) (59.9%). It covers short factual questions. The board median is 41.0%, so Muse Spark 1.2 (xhigh) is 19.3 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Muse Spark 1.2 (xhigh) manages 60.3% on WeirdML, good for #19 of 109. GPT-6 Astra (max) leads at 93.3%, 33.0 points further up, and 83% of models sit below. The results either side are GPT-5 (high) (60.7%) and GLM-5.1 (57.1%). It covers unusual machine-learning tasks. The board median is 41.2%, so Muse Spark 1.2 (xhigh) is 19.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Muse Spark 1.2 (xhigh) lands below the middle on DeepSWE: 54.9%, #15 of 19, with 74% of the field scoring higher. Gemini 3.8 Flash (high) leads at 73.8%; the closest result above is Qwen3.8 Max (xhigh) at 57.5%. It covers long-horizon software engineering. The board median is 67.0%, so Muse Spark 1.2 (xhigh) is 12.2 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Muse Spark 1.2 (xhigh) lands below the middle on GDP.pdf: 16.0%, #20 of 25, with 76% of the field scoring higher. GPT-6 Astra (max) leads at 34.2%; the closest result above is Qwen3.8 Flash (xhigh) at 16.6%. It covers economically valuable document tasks. The board median is 22.8%, so Muse Spark 1.2 (xhigh) is 6.8 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
LMArena Code Elo is a weak spot. Muse Spark 1.2 (xhigh) scores 1,532, #25 of 30, behind 80% of the models measured. The leader, Claude Opus 5.5 (max), reaches 1,814. It covers crowd-voted coding answers. The board median is 1,622, so Muse Spark 1.2 (xhigh) is 90 points below it. Source: LMArena, captured 2026-10-06.
FrontierSWE is a weak spot. Muse Spark 1.2 (xhigh) scores 12.0%, #18 of 19, behind 89% of the models measured. The leader, GPT-6 Astra (max), reaches 65.5%. It covers hard multi-hour software engineering. The board median is 29.5%, so Muse Spark 1.2 (xhigh) is 17.5 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
The full table above also lists the other 3 results, including boards with fewer than ten models, where a rank says little.
Strengths and weaknesses
Software engineering: uneven, ahead of 40% of the field on average across 5 benchmarks. Its best board here is SciCode (#19 of 106) and its weakest is FrontierSWE (#18 of 19). Agents and tool use: solid, ahead of 56% of the field on average across 2 benchmarks. Its best board here is DTBench (#11 of 146) and its weakest is GDP.pdf (#20 of 25). General knowledge and chat: strong, ahead of 85% of the field on average across 5 benchmarks. Its best board here is LMArena Vision Elo (#3 of 22) and its weakest is WeirdML (#19 of 109).
Price and value
The median input price across the 427 priced models is $0.500 per million tokens, so Muse Spark 1.2 (xhigh) costs more than the typical listing. Models that score within 1.5 points or better and cost less: Qwen3.7 Flash ($0.030, 50.5), GLM-5.2 (high) ($0.060, 51.6), glm-4.7-flash ($0.060, 50.1). Similar scores cost more on glm-5p1 ($1.40) and gemini-3.5-flash-high ($1.50).
Release history
The Meta AI model released before it was Muse Spark 1.1 (xhigh) on 2026-07-09, which scores 52.4. The next one was Muse Glimmer (high) on 2026-08-10, at 48.3.
Reading these numbers
The 13 scores on this page come from 2 sources: Epoch AI Benchmarking Hub (10), LMArena (3). No official leaderboard supplied any of them, so treat them as third-party measurements. The leaderboard also lists 1 other reasoning-effort setting of this model, which redirect here. This page shows the highest.
Muse Spark 1.2 (xhigh) is also listed at 1 other reasoning-effort setting; this page shows the highest-effort row, which is the one the leaderboard keeps by default.
Models ranked nearby
- #110 Hy3-preview (51.6)
- #111 LongCat-Flash-Chat (51.6)
- #112 Darwin-27B-Opus (51.5)
- #114 NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 (51.5)
- #115 glm-5p1 (51.5)
- #116 Ornith-1.5-9B (51.5)
More from Meta AI
- Muse Spark 1.3 (max) (#18, 55.1)
- Muse Spark (#29, 53.7)
- Muse Spark 1.1 (xhigh) (#66, 52.4)
- Llama 3.1-405B (#142, 51.2)
- Muse Wonder (max) (#209, 50.1)
- Muse Glimmer (high) (#361, 48.3)
Compare it on the live leaderboard → · How the aggregate score works