Gemini 4 Carbon vs Muse Ember
TryTop 2 now
Gemini 4 Carbon leads by 5.5 capability points (≈95% interval +2.9 to +8.2): a clear lead over Muse Ember · P(higher) ≈ 100%
Wins 54 of 61 shared benchmarks (7 lost)
Capability vs. relative cost
Gemini 4 CarbonMuse EmberClosed weightsOpen weights
Benchmark wins54 – 7 · 61 shared
Comparing matching setups only (3 benchmarks excluded) · suites counted once
Area← Gemini 4 CarbonMuse Ember →
Data: Artificial Analysis, Vals AI, Surge AI and 10 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Ember minus Gemini 4 Carbon.
| FoodTruck Bench | Business simulations | 77218 | 34639 | −54.7 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 69.1 | 25.1 | −44.0 |
| Terminal-Bench-Science 0.1 | Science › Multidisciplinary | 62.4 | 20.0 | −42.4 |
| FrontierCUA v1 pass@1 | Tool & computer use › Computer & browser use | 42.5 | 7.5 | −35.0 |
| Code Migration v1.1 (Vals Index v2.2 subset) | Software › Software engineering | 80.1 | 47.1 | −32.9 |
| DAXBench v2 | Software › Data & ML | 93.6 | 62.8 | −30.9 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 16.1 | 1.0 | −29.5 |
| RiemannBench (November 2026 rescoring) | Math | 80.0 | 53.6 | −26.4 |
| BuseyBench SVG | Perception & media › Generation | 7.9 | 5.4 | −24.6 |
| Artificial Analysis Intelligence Index v5.0.1 | Composite indexes | 60.0 | 36.0 | −24.0 |
| Terminal-Bench 4.0 | Software › Software engineering | 63.1 | 39.4 | −23.7 |
| HiL-Bench (Human-in-Loop Benchmark) | Tool & computer use › General agents | 50.3 | 74.0 | +23.7 |
| Ramp SWE-Bench v2 | Software › Software engineering | 75.4 | 51.8 | −23.6 |
| Artificial Analysis Engineering Index | Science › Engineering | 65.3 | 41.8 | −23.5 |
| GDP.xlsx | Professional › General knowledge work | 38.9 | 15.4 | −23.5 |
| Artificial Analysis Legal Index | Professional › Legal | 62.0 | 38.5 | −23.5 |
| AA-Omniscience v1.1 Index | Knowledge & reasoning › General knowledge | 53.5 | 7.7 | −22.9 |
| Terminal-Bench 4.0 (Artificial Analysis) | Software › Software engineering | 62.6 | 40.9 | −21.7 |
| LMArena Code Arena WebDev | Software › Frontend & design | 1769 | 1608 | −21.3 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 39.4 | -2.7 | −21.1 |
| HANDBOOK.md Agents v1.1 | Tool & computer use › Customer service & operations | 49.8 | 29.8 | −20.0 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 82.3 | 62.3 | −20.0 |
| SimpleBench | Knowledge & reasoning › General knowledge | 92.6 | 72.7 | −19.9 |
| DrugDiscoveryBench | Science › Life sciences | 71.5 | 52.4 | −19.1 |
| Artificial Analysis Economics Index | Composite indexes | 59.6 | 41.0 | −18.7 |
| EnterpriseBench: CoreCraft | Tool & computer use › Customer service & operations | 80.5 | 62.1 | −18.4 |
| Artificial Analysis Finance & Accounting Index | Professional › Finance, accounting & tax | 62.6 | 44.6 | −18.0 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 35.2 | 17.2 | −18.0 |
| Artificial Analysis Coding Agent Index v2.1 | Software › Software engineering | 55.5 | 39.7 | −15.8 |
| Artificial Analysis Healthcare & Medical Index | Health & medicine | 56.8 | 41.3 | −15.5 |
| SWE Atlas v1.1 - Overall | Software › Software engineering | 71.5 | 56.0 | −15.5 |
| APEX-Agents 1.1 | Professional › General knowledge work | 90.4 | 75.7 | −14.7 |
| Artificial Analysis Coding Agent Index v2.0 | Software › Software engineering | 59.4 | 45.9 | −13.5 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 40.6 | 27.2 | −13.4 |
| GBENCH One-shot Coding GScore | Games | 69.0 | 56.4 | −12.6 |
| Artificial Analysis Strategy & Ops Index | Professional › General knowledge work | 60.9 | 48.7 | −12.3 |
| ComplexConstraints (November 2026 rescoring) | Language & writing › Instruction following | 70.9 | 58.8 | −12.1 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 77.8 | 68.1 | −9.7 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 17.2 | -2.0 | −9.6 |
| Artificial Analysis Progress Index v1 | Composite indexes | 351 | 329 | −7.7 |
| HLAB-Vals-1.1 | Professional › Legal | 17.5 | 25.0 | +7.5 |
| GDP.pdf (November 2026 rescoring) | Professional › General knowledge work | 56.4 | 49.2 | −7.2 |
| GDP.pdf (Artificial Analysis, All-pass) | Professional › General knowledge work | 29.6 | 22.8 | −6.8 |
| Legal Research Bench | Professional › Legal | 55.8 | 49.0 | −6.7 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 17.9 | 4.8 | −6.6 |
| Chartography (November 2026 rescoring) | Perception & media › Documents & charts | 85.5 | 79.3 | −6.2 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | -5.7 | 6.6 | +6.2 |
| Vals Finance Index | Professional › Finance, accounting & tax | 67.0 | 60.9 | −6.1 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 89.1 | 83.7 | −5.4 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 76.4 | 71.4 | −5.1 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1611 | 1526 | −4.6 |
| FORTRESS Average Risk Score (ARS) | Safety & alignment | 16.4 | 11.9 | +4.5 |
| Excel Modeling Benchmark v1.1 | Professional › Finance, accounting & tax | 76.5 | 72.2 | −4.3 |
| Finance Agent v2.1 | Professional › Finance, accounting & tax | 67.2 | 63.8 | −3.3 |
| τ³-Banking (Mercor run, bm25_grep retrieval) | Tool & computer use › Customer service & operations | 55.7 | 53.3 | −2.4 |
| Professional Reasoning Benchmark - Legal | Professional › Legal | 53.4 | 55.6 | +2.3 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 19.4 | 18.3 | −1.1 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 23.8 | 24.2 | +0.4 |
| Harvey LAB (Mercor run, undisclosed ~213-task cohort) | Professional › Legal | 23.0 | 22.7 | −0.3 |
| Vals Legal Index | Professional › Legal | 75.9 | 75.7 | −0.2 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 0.8 | 0.9 | +0.1 |
Data: Artificial Analysis, Vals AI, Surge AI and 10 othersatlas.village.dev