- Artificial Analysis Intelligence Index 39
- JobBench 61.8
- MCP-Atlas 89.2
- GDPVal-AA v2 1401
- OSWorld Verified 80.4
- Toolathlon Verified 72.8
- HLE-Diamond (no tools) 26.1
- AA Omniscience 25
- SciCode 58.5
- HLE (w/ tools) 61
- GPQA Diamond 89
- DeepSWE 1.1 Preferred harness 60
- SWE-Atlas-QnA Preferred harness 58
- Terminal-Bench 2.1 Terminus 2 agent 76.4
- ZeroBench (w/ tools) 50
- BabyVision (w/ tools) 80.2
- SimpleVQA (w/ tools) 73.6
- WorldVQA (w/ tools) 49.4
- IntPhys2 68.3
- VideoMMEv2 (subtitle setting not stated) 61.5
- FORTRESS adversarial (static), ARS 5.8 / ORS 17.4, lower is better 5.8
- StrongREJECT v2 (adaptive), lower is better 0.5
- MASK honesty rate (text) 98.1
- Agentic Misalignment benchmark, misbehavior rate 0% (text, lower is better) 0
Muse Spark Open
Muse Spark Open ranks first on 3 of the 63 benchmarks it has scores on.
Meta Released 15 Oct 2026 66
Benchmarks with a score
3
Ranked first, out of five or more models
10
In the top three
Meta Oct 2026 Muse Spark Open
Open weights
155.1
#39 of 271
Epoch Capabilities Index
- SciCode 58.5% #5 /47
- MultiChallenge 74.2% #6 /35
- SimpleVQA 73.6% #5 /23
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Humanity's Last Exam | 61% | #11 of 147 | Model card | |
| MultiChallenge | 74.2% | #6 of 35 | Official leaderboard | |
| SimpleVQA | 73.6% | #5 of 23 | Model card | |
| SimpleQA Verified | 56.6% | #20 of 84 | Epoch AI | |
| HLE-Diamond | 28.2% | #15 of 18 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.1% | #30 of 202 | Epoch AI | |
| GPQA Diamond | 88.6% | #51 of 219 | Epoch AI | |
| CritPt | 14.9% | #46 of 160 | Epoch AI | |
| FrontierMath Tiers 1-3 v2 | 69.8% | #29 of 89 | Epoch AI | |
| ARC-AGI | 89% | #37 of 92 | Official leaderboard | |
| ARC-AGI-2 | 57.5% | #39 of 95 | Official leaderboard | |
| FrontierMath Tier 4 v2 | 31.7% | #33 of 71 | Epoch AI | |
| BrokenArXiv | 33% | #5 of 6 | Official leaderboard | |
| ArXivMath | 43.5% | #6 of 6 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SciCode | 58.5% | #5 of 47 | Model card | |
| ALE-Bench | 1,590 | #20 of 121 | Epoch AI | |
| Vibe Code Bench | 76.2% | #38 of 115 | Official leaderboard | |
| Code Arena | 1,526 | #47 of 124 | Official leaderboard | |
| SWE Atlas | 58% | #5 of 11 | Model card | |
| Code Migration | 21.2% | #51 of 78 | Official leaderboard | |
| DeepSWE | 60% | #34 of 51 | Model card | |
| SWE-Bench Pro V2 | 95% | #11 of 16 | Official leaderboard | |
| SWE-Bench Pro V2 Hard | 82.3% | #11 of 16 | Official leaderboard | |
| Code Migration v1.1 | 26.9% | #24 of 29 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MCP Atlas | 89.2% | #1 of 40 | Model card | |
| OSWorld-Verified | 80.4% | #10 of 37 | Model card | |
| Terminal-Bench 2.1 | 76.4% | #29 of 53 | Model card | |
| Toolathlon-Verified | 72.8% | #10 of 13 | Model card | |
| ARC-AGI-3 | 0.42% | #19 of 24 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Vals Finance Agent v2 | 61.3% | #3 of 80 | Official leaderboard | |
| PRBench | 59.8% | #2 of 44 | Official leaderboard | |
| Harvey LAB v1.1 (HLAB-Vals-1.1) | 25% | #2 of 33 | Official leaderboard | |
| Harvey LAB (Legal Agent Benchmark) | 16.7% | #8 of 80 | Official leaderboard | |
| FrontierFinance | 53.7% | #2 of 16 | Official leaderboard | |
| JobBench | 61.8% | #5 of 12 | Model card | |
| Legal Research Bench | 38.5% | #38 of 88 | Official leaderboard | |
| Tax Agent Bench | 26.3% | #38 of 79 | Official leaderboard | |
| Vals Finance Agent v2.1 | 60.7% | #19 of 35 | Official leaderboard | |
| Vals Legal Index | 68.5% | #20 of 33 | Official leaderboard | |
| GDPval-AA v2 | 1,401 | #17 of 26 | Model card | |
| Excel Modeling Benchmark | 53.7% | #51 of 76 | Official leaderboard | |
| Vals Finance Index | 53% | #28 of 35 | Official leaderboard | |
| Vals Index v2.2 | 48.6% | #27 of 32 | Official leaderboard | |
| Excel Modeling Benchmark v1.1 | 62% | #32 of 37 | Official leaderboard | |
| APEX-Agents | 33.4% | #49 of 56 | Epoch AI |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ZeroBench | 50% | #2 of 25 | Model card | |
| BabyVision | 80.2% | #8 of 20 | Model card | |
| Humanity's Sixth Sense | 29.6% | #19 of 29 | Official leaderboard | |
| VideoMME v2 (with subtitles) | 61.5% | #4 of 6 | Model card | |
| WorldVQA | 49.4% | #8 of 9 | Model card | |
| IntPhys 2 | 68.3% | #1 of 1 | Model card |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| DistressBench | 87.7% | #2 of 42 | Official leaderboard | |
| MASK | 98.1% | #1 of 21 | Model card | |
| FORTRESS | 5.80% | #1 of 6 | Model card | |
| PCT (Polarized Contrastive Pairs) | 41.3% | #18 of 46 | Official leaderboard | |
| Agentic misalignment suite (Anthropic) | 0% | #3 of 7 | Model card | |
| BioSecBench-Refusal | 34.9% | #11 of 22 | Official leaderboard | |
| StrongREJECT v2 (Meta) | 0.50% | #2 of 4 | Model card | |
| OncologyBench | 12% | #24 of 30 | Official leaderboard | |
| CheatBench | 84.2% | #10 of 10 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| CAIS Risk Index | 39.1 | #12 of 65 | Official leaderboard | |
| CAIS Vision Capabilities Index | 59.4 | #17 of 53 | Official leaderboard | |
| CAIS Text Capabilities Index | 36.3 | #28 of 66 | Official leaderboard | |
| SAGE v1.1 (Vals) | 85.8% | #13 of 28 | Official leaderboard | |
| Artificial Analysis Intelligence Index v4.3 | 39.0 | #3 of 4 | Model card | |
| LiveBench | 51.6% | #25 of 28 | Official leaderboard |
Meta's models over time
Meta's models in color, against every other model on the Epoch Capabilities Index, by release date.
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.