GPT-6.2 Sol vs Muse Ember
GPT-6.2 Sol leads by 4.8 capability points (≈95% interval +2.2 to +7.4): a clear lead over Muse Ember · P(higher) ≈ 100%
Wins 37 of 49 shared benchmarks (11 lost, 1 tied)
Capability vs. relative cost
GPT-6.2 SolMuse EmberClosed weightsOpen weights
Benchmark wins37 – 11 · 49 shared, 1 tied
Comparing matching setups only (8 benchmarks excluded) · suites counted once
Area← GPT-6.2 SolMuse Ember →
Data: Artificial Analysis, Vals AI, Surge AI and 5 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Ember minus GPT-6.2 Sol.
| Terminal-Bench-Science 0.1 | Science › Multidisciplinary | 67.6 | 20.0 | −47.6 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 64.0 | 25.1 | −38.9 |
| Code Migration v1.1 (Vals Index v2.2 subset) | Software › Software engineering | 85.3 | 47.1 | −38.2 |
| TaxCalcBench TY25 | Professional › Finance, accounting & tax | 32.0 | 2.0 | −30.0 |
| RiemannBench (November 2026 rescoring) | Math | 80.0 | 53.6 | −26.4 |
| EnterpriseBench: CoreCraft | Tool & computer use › Customer service & operations | 36.4 | 62.1 | +25.7 |
| Artificial Analysis Intelligence Index v5.0.1 | Composite indexes | 57.3 | 36.0 | −21.2 |
| Terminal-Bench 4.0 | Software › Software engineering | 60.6 | 39.4 | −21.2 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 11.3 | 1.0 | −20.2 |
| Artificial Analysis Engineering Index | Science › Engineering | 61.2 | 41.8 | −19.4 |
| Terminal-Bench 4.0 (Artificial Analysis) | Software › Software engineering | 60.1 | 40.9 | −19.2 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 37.2 | 18.3 | −18.9 |
| Ramp SWE-Bench v2 (Bash harness) | Software › Software engineering | 72.4 | 54.4 | −18.0 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 44.4 | 27.2 | −17.2 |
| Artificial Analysis Legal Index | Professional › Legal | 55.0 | 38.5 | −16.6 |
| Artificial Analysis Economics Index | Composite indexes | 57.4 | 41.0 | −16.4 |
| AA-Omniscience v1.1 Index | Knowledge & reasoning › General knowledge | 40.4 | 7.7 | −16.3 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 78.6 | 62.3 | −16.3 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 23.3 | -2.7 | −13.0 |
| HLAB-Vals-1.1 | Professional › Legal | 12.1 | 25.0 | +12.9 |
| Artificial Analysis Finance & Accounting Index | Professional › Finance, accounting & tax | 57.2 | 44.6 | −12.6 |
| Professional Reasoning Benchmark - Legal | Professional › Legal | 43.3 | 55.6 | +12.3 |
| Legal Research Bench | Professional › Legal | 37.0 | 49.0 | +12.0 |
| Artificial Analysis Healthcare & Medical Index | Health & medicine | 53.3 | 41.3 | −12.0 |
| Artificial Analysis Strategy & Ops Index | Professional › General knowledge work | 59.4 | 48.7 | −10.7 |
| ComplexConstraints (November 2026 rescoring) | Language & writing › Instruction following | 69.5 | 58.8 | −10.7 |
| DrugDiscoveryBench | Science › Life sciences | 62.2 | 52.4 | −9.8 |
| Vals Legal Index | Professional › Legal | 66.7 | 75.7 | +9.0 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 32.8 | 24.2 | −8.6 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 62.9 | 71.4 | +8.5 |
| Chartography (November 2026 rescoring) | Perception & media › Documents & charts | 87.7 | 79.3 | −8.4 |
| GDP.pdf (Artificial Analysis, All-pass) | Professional › General knowledge work | 30.0 | 22.8 | −7.2 |
| GDP.xlsx | Professional › General knowledge work | 22.3 | 15.4 | −6.9 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 10.9 | -2.0 | −6.5 |
| Vals Index v2.2 | Composite indexes | 66.1 | 60.5 | −5.6 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1612 | 1520 | −5.0 |
| Finance Agent v2.1 | Professional › Finance, accounting & tax | 58.9 | 63.8 | +4.9 |
| HANDBOOK.md Agents v1.1 | Tool & computer use › Customer service & operations | 34.5 | 29.8 | −4.7 |
| SWE Atlas v1.1 - Overall | Software › Software engineering | 60.5 | 56.0 | −4.5 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 13.4 | 4.8 | −4.3 |
| Vals Finance Index | Professional › Finance, accounting & tax | 56.7 | 60.9 | +4.3 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 87.6 | 83.7 | −3.9 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 71.3 | 68.1 | −3.2 |
| SimpleBench | Knowledge & reasoning › General knowledge | 69.9 | 72.7 | +2.8 |
| Excel Modeling Benchmark v1.1 | Professional › Finance, accounting & tax | 74.3 | 72.2 | −2.1 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | 4.3 | 6.6 | +1.2 |
| GDP.pdf (November 2026 rescoring) | Professional › General knowledge work | 48.8 | 49.2 | +0.4 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 1.2 | 0.9 | −0.1 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 17.2 | 17.2 | 0.0 |
Data: Artificial Analysis, Vals AI, Surge AI and 5 othersatlas.village.dev