Gemini 4 Neon vs GPT-6.1 Terra
Gemini 4 Neon leads by 3.7 capability points (≈95% interval +0.9 to +6.6): a clear lead over GPT-6.1 Terra · P(higher) ≈ 100%
Wins 47 of 50 shared benchmarks (2 lost, 1 tied)
Capability vs. relative cost
Gemini 4 NeonGPT-6.1 TerraClosed weightsOpen weights
Benchmark wins47 – 2 · 50 shared, 1 tied
Comparing matching setups only (31 benchmarks excluded) · suites counted once
Area← Gemini 4 NeonGPT-6.1 Terra →
Data: Artificial Analysis, Vals AI, Epoch AI and 6 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6.1 Terra minus Gemini 4 Neon.
| Vending-Bench 2 | Business simulations | 16843 | 9107 | −45.8 |
| Legal Research Bench | Professional › Legal | 56.3 | 26.0 | −30.3 |
| Terminal-Bench-Science 0.1 | Science › Multidisciplinary | 70.0 | 41.9 | −28.1 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 79.8 | 55.0 | −24.8 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 71.3 | 47.3 | −24.0 |
| τ³-Banking (Mercor run, bm25_grep retrieval) | Tool & computer use › Customer service & operations | 59.5 | 37.1 | −22.4 |
| SimpleBench | Knowledge & reasoning › General knowledge | 94.4 | 72.2 | −22.2 |
| Harvey LAB (Mercor run, undisclosed ~213-task cohort) | Professional › Legal | 26.6 | 4.5 | −22.1 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1699 | 1334 | −20.0 |
| Vals Legal Index | Professional › Legal | 78.4 | 58.8 | −19.6 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 79.3 | 60.2 | −19.2 |
| HLAB-Vals-1.1 | Professional › Legal | 19.6 | 1.7 | −17.9 |
| SimpleQA Verified (Epoch anti-abstention) | Knowledge & reasoning › General knowledge | 75.2 | 57.3 | −17.9 |
| Artificial Analysis Legal Index | Professional › Legal | 63.8 | 46.4 | −17.4 |
| Terminal-Bench 4.0 (Artificial Analysis) | Software › Software engineering | 63.6 | 46.5 | −17.2 |
| Artificial Analysis Finance & Accounting Index | Professional › Finance, accounting & tax | 65.1 | 49.0 | −16.1 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 18.4 | 10.5 | −15.5 |
| Artificial Analysis Healthcare & Medical Index | Health & medicine | 60.1 | 44.7 | −15.4 |
| Vals Finance Index | Professional › Finance, accounting & tax | 68.3 | 53.3 | −15.0 |
| Artificial Analysis Strategy & Ops Index | Professional › General knowledge work | 63.8 | 50.0 | −13.8 |
| Artificial Analysis Economics Index | Composite indexes | 61.9 | 48.5 | −13.4 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 44.5 | 17.9 | −13.3 |
| MKA (Artificial Analysis) | Professional › General knowledge work | 56.5 | 43.8 | −12.7 |
| Finance Agent v2.1 | Professional › Finance, accounting & tax | 68.6 | 56.0 | −12.6 |
| AA-Omniscience v1.1 Index | Knowledge & reasoning › General knowledge | 54.8 | 29.6 | −12.6 |
| Terminal-Bench 4.0 | Software › Software engineering | 61.6 | 49.5 | −12.1 |
| Artificial Analysis Engineering Index | Science › Engineering | 66.8 | 54.7 | −12.1 |
| Mystery Game Puzzles | Games | 85.0 | 74.0 | −11.0 |
| Vals Index v2.2 | Composite indexes | 70.9 | 60.5 | −10.3 |
| Chartography | Perception & media › Documents & charts | 73.5 | 63.2 | −10.3 |
| Artificial Analysis Intelligence Index v5.0.1 | Composite indexes | 61.4 | 51.3 | −10.1 |
| Furniture Assembly Benchmark | Perception & media › Spatial & visual reasoning | 89.2 | 79.2 | −10.0 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 93.7 | 84.0 | −9.7 |
| APEX-Agents 1.1 | Professional › General knowledge work | 88.9 | 79.3 | −9.6 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | -6.0 | 12.7 | +9.3 |
| LiveBench 2.0 (2026-11-06 release) | Composite indexes | 87.4 | 78.6 | −8.9 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 44.4 | 36.9 | −7.6 |
| APEX-Accounting 1.1 | Professional › Finance, accounting & tax | 70.8 | 63.7 | −7.1 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 19.2 | 8.5 | −5.3 |
| FrontierMath Tier 4 (v2) - Private | Math | 97.6 | 92.7 | −4.9 |
| GDP.pdf (Artificial Analysis, All-pass) | Professional › General knowledge work | 33.4 | 28.6 | −4.8 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 21.2 | 13.2 | −4.0 |
| Code Migration v1.1 (Vals Index v2.2 subset) | Software › Software engineering | 75.8 | 72.2 | −3.6 |
| Excel Modeling Benchmark v1.1 | Professional › Finance, accounting & tax | 76.5 | 73.3 | −3.1 |
| GPQA Diamond | Science › Multidisciplinary | 96.3 | 94.3 | −2.0 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 31.7 | 30.6 | −1.1 |
| FrontierMath Tiers 1-3 (v2) - Private | Math | 91.2 | 90.9 | −0.4 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 77.4 | 77.1 | −0.3 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 1.0 | 1.2 | +0.1 |
| OTIS Mock AIME 2024-2025 | Math | 100.0 | 100.0 | 0.0 |
Data: Artificial Analysis, Vals AI, Epoch AI and 6 othersatlas.village.dev