GPT-6.1 Terra vs Muse Ember
GPT-6.1 Terra leads by 3.1 capability points (≈95% interval +0.6 to +5.5): a clear lead over Muse Ember · P(higher) ≈ 99%
Wins 36 of 51 shared benchmarks (15 lost)
Capability vs. relative cost
GPT-6.1 TerraMuse EmberClosed weightsOpen weights
Benchmark wins36 – 15 · 51 shared
Comparing matching setups only (9 benchmarks excluded) · suites counted once
Area← GPT-6.1 TerraMuse Ember →
Data: Artificial Analysis, Vals AI, Arena and 10 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Ember minus GPT-6.1 Terra.
| Terminal-Bench-Science 0.1 | Science › Multidisciplinary | 57.6 | 20.0 | −37.6 |
| DAXBench v2 | Software › Data & ML | 93.6 | 62.8 | −30.9 |
| Code Migration v1.1 (Vals Index v2.2 subset) | Software › Software engineering | 72.2 | 47.1 | −25.1 |
| HLAB-Vals-1.1 | Professional › Legal | 1.7 | 25.0 | +23.3 |
| Legal Research Bench | Professional › Legal | 26.0 | 49.0 | +23.1 |
| KradleBench | Games | 48.6 | 25.7 | −22.9 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 47.3 | 25.1 | −22.2 |
| TaxCalcBench TY25 | Professional › Finance, accounting & tax | 24.0 | 2.0 | −22.0 |
| GBENCH One-shot Coding GScore | Games | 76.4 | 56.4 | −20.0 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 81.4 | 62.3 | −19.1 |
| BuseyBench SVG | Perception & media › Generation | 7.3 | 5.4 | −18.8 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 10.5 | 1.0 | −18.6 |
| Harvey LAB (Mercor run, undisclosed ~213-task cohort) | Professional › Legal | 4.5 | 22.7 | +18.2 |
| Vals Legal Index | Professional › Legal | 58.8 | 75.7 | +16.9 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 55.0 | 71.4 | +16.3 |
| τ³-Banking (Mercor run, bm25_grep retrieval) | Tool & computer use › Customer service & operations | 37.1 | 53.3 | +16.2 |
| Artificial Analysis Intelligence Index v5.0.1 | Composite indexes | 51.3 | 36.0 | −15.3 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 42.3 | 27.2 | −15.2 |
| Terminal-Bench 4.0 (Artificial Analysis) | Software › Software engineering | 55.6 | 40.9 | −14.6 |
| Artificial Analysis Engineering Index | Science › Engineering | 54.7 | 41.8 | −13.0 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 30.6 | 18.3 | −12.2 |
| AA-Omniscience v1.1 Index | Knowledge & reasoning › General knowledge | 29.6 | 7.7 | −10.9 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 17.9 | -2.7 | −10.3 |
| Terminal-Bench 4.0 | Software › Software engineering | 49.5 | 39.4 | −10.1 |
| Artificial Analysis Legal Index | Professional › Legal | 46.4 | 38.5 | −7.9 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 60.2 | 68.1 | +7.9 |
| Finance Agent v2.1 | Professional › Finance, accounting & tax | 56.0 | 63.8 | +7.9 |
| Vals Finance Index | Professional › Finance, accounting & tax | 53.3 | 60.9 | +7.6 |
| Artificial Analysis Economics Index | Composite indexes | 48.5 | 41.0 | −7.5 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1400 | 1520 | +6.6 |
| FORTRESS Average Risk Score (ARS) | Safety & alignment | 18.2 | 11.9 | +6.3 |
| APEX-1 (Mercor, mean score) | Professional › General knowledge work | 64.2 | 70.4 | +6.2 |
| Ramp SWE-Bench v2 (Bash harness) | Software › Software engineering | 60.6 | 54.4 | −6.2 |
| HiL-Bench (Human-in-Loop Benchmark) | Tool & computer use › General agents | 68.0 | 74.0 | +6.0 |
| GDP.pdf (Artificial Analysis, All-pass) | Professional › General knowledge work | 28.6 | 22.8 | −5.8 |
| Artificial Analysis Progress Index v1 | Composite indexes | 345 | 329 | −5.8 |
| DrugDiscoveryBench | Science › Life sciences | 58.1 | 52.4 | −5.7 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 8.5 | -2.0 | −5.3 |
| Artificial Analysis Finance & Accounting Index | Professional › Finance, accounting & tax | 49.0 | 44.6 | −4.4 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 13.2 | 4.8 | −4.2 |
| APEX-Agents 1.1 | Professional › General knowledge work | 79.3 | 75.7 | −3.6 |
| Artificial Analysis Healthcare & Medical Index | Health & medicine | 44.7 | 41.3 | −3.4 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | 12.7 | 6.6 | −3.0 |
| SWE Atlas v1.1 - Overall | Software › Software engineering | 57.7 | 56.0 | −1.7 |
| Artificial Analysis Strategy & Ops Index | Professional › General knowledge work | 50.0 | 48.7 | −1.3 |
| Excel Modeling Benchmark v1.1 | Professional › Finance, accounting & tax | 73.3 | 72.2 | −1.2 |
| SimpleBench | Knowledge & reasoning › General knowledge | 72.2 | 72.7 | +0.5 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 84.0 | 83.7 | −0.3 |
| ParseBench-2 | Perception & media › Documents & charts | 74.3 | 74.5 | +0.3 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 1.2 | 0.9 | −0.1 |
| Vals Index v2.2 | Composite indexes | 60.5 | 60.5 | −0.1 |
Data: Artificial Analysis, Vals AI, Arena and 10 othersatlas.village.dev