Gemini 4 Neon vs Grok 4.7
Gemini 4 Neon leads by 8.4 capability points (≈95% interval +5.7 to +11.0): a clear lead over Grok 4.7 · P(higher) ≈ 100%
Wins 70 of 72 shared benchmarks (2 lost)
Capability vs. relative cost
Gemini 4 NeonGrok 4.7Closed weightsOpen weights
Benchmark wins70 – 2 · 72 shared
Comparing matching setups only (2 benchmarks excluded) · suites counted once
Area← Gemini 4 NeonGrok 4.7 →
Data: Artificial Analysis, Vals AI, Surge AI and 9 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Grok 4.7 minus Gemini 4 Neon.
| FrontierMath Tier 4 (v2) - Private | Math | 97.6 | 17.1 | −80.5 |
| Furniture Assembly Benchmark | Perception & media › Spatial & visual reasoning | 89.2 | 20.8 | −68.3 |
| Chartography | Perception & media › Documents & charts | 73.5 | 14.7 | −58.8 |
| Terminal-Bench-Science 0.1 | Science › Multidisciplinary | 70.0 | 11.4 | −58.6 |
| Mystery Game Puzzles | Games | 85.0 | 29.0 | −56.0 |
| Terminal-Bench Science 0.1 (Vals) | Science › Multidisciplinary | 62.9 | 10.0 | −52.9 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 71.3 | 23.6 | −47.7 |
| Chess Puzzles (task version 2.0) | Games | 83.0 | 41.0 | −42.0 |
| FrontierMath Tiers 1-3 (v2) - Private | Math | 91.2 | 53.0 | −38.2 |
| Terminal-Bench 4.0 (Artificial Analysis) | Software › Software engineering | 63.6 | 25.8 | −37.9 |
| Vending-Bench 2 | Business simulations | 16843 | 10537 | −37.4 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 18.4 | 1.5 | −33.0 |
| Terminal-Bench 4.0 | Software › Software engineering | 61.6 | 28.8 | −32.8 |
| Bug Hunt Bench | Software › Software engineering | 60.0 | 27.6 | −32.4 |
| DAYJOB: Accounting | Professional › Finance, accounting & tax | 49.6 | 20.0 | −29.6 |
| Vals Science Index | Science › Multidisciplinary | 62.6 | 34.8 | −27.8 |
| τ³-Banking (Mercor run, bm25_grep retrieval) | Tool & computer use › Customer service & operations | 59.5 | 32.6 | −26.9 |
| RiemannBench (November 2026 rescoring) | Math | 84.0 | 57.6 | −26.4 |
| sudo L7 (October 2026 rescoring) | Software › Software engineering | 38.3 | 12.2 | −26.1 |
| Artificial Analysis Intelligence Index v5.0.1 | Composite indexes | 61.4 | 35.5 | −25.9 |
| Artificial Analysis Engineering Index | Science › Engineering | 66.8 | 41.8 | −25.1 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 50.5 | 26.5 | −24.0 |
| Code Migration v1.1 (Vals Index v2.2 subset) | Software › Software engineering | 75.8 | 51.8 | −23.9 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 44.5 | 2.1 | −21.2 |
| GDP.xlsx | Professional › General knowledge work | 43.1 | 22.3 | −20.8 |
| MysteryMechanism | Math | 45.0 | 25.2 | −19.8 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 47.2 | 27.6 | −19.6 |
| LiveBench 2.0 (2026-11-06 release) | Composite indexes | 87.4 | 67.9 | −19.5 |
| SimpleQA Verified (Epoch anti-abstention) | Knowledge & reasoning › General knowledge | 75.2 | 56.0 | −19.2 |
| Artificial Analysis Coding Agent Index v2.1 | Software › Software engineering | 58.6 | 41.6 | −17.0 |
| CursorBench 4.1 | Software › Software engineering | 65.4 | 48.6 | −16.8 |
| LMArena Code Arena WebDev | Software › Frontend & design | 1764 | 1639 | −16.6 |
| Artificial Analysis Legal Index | Professional › Legal | 63.8 | 47.2 | −16.6 |
| MKA (Artificial Analysis) | Professional › General knowledge work | 56.5 | 40.1 | −16.3 |
| Artificial Analysis Economics Index | Composite indexes | 61.9 | 45.7 | −16.2 |
| CWE-bench v2 Pass@1 | Software › Cybersecurity | 70.7 | 54.7 | −16.0 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 44.4 | 28.7 | −15.8 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 78.0 | 62.3 | −15.7 |
| APEX-Accounting 1.1 | Professional › Finance, accounting & tax | 70.5 | 55.3 | −15.2 |
| Vals RSI Index v1.1 (four tasks) | Software › Data & ML | 34.4 | 20.2 | −14.2 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 79.3 | 65.6 | −13.8 |
| Artificial Analysis Healthcare & Medical Index | Health & medicine | 60.1 | 46.6 | −13.5 |
| Artificial Analysis Finance & Accounting Index | Professional › Finance, accounting & tax | 65.1 | 51.7 | −13.4 |
| LMArena Text Arena Elo | Language & writing › General assistance | 1518 | 1443 | −13.2 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 21.2 | -4.3 | −12.7 |
| Vals Index v2.2 | Composite indexes | 70.9 | 58.6 | −12.3 |
| ComplexConstraints (November 2026 rescoring) | Language & writing › Instruction following | 72.9 | 60.8 | −12.1 |
| AA-Omniscience v1.1 Index | Knowledge & reasoning › General knowledge | 54.8 | 31.2 | −11.8 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 79.8 | 69.0 | −10.8 |
| BioMysteryBench (Vals Terminus 2) | Science › Life sciences | 80.0 | 69.3 | −10.7 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 31.7 | 21.1 | −10.6 |
| GDP.pdf (Artificial Analysis, All-pass) | Professional › General knowledge work | 33.4 | 23.2 | −10.2 |
| Vals Finance Index | Professional › Finance, accounting & tax | 68.3 | 58.8 | −9.5 |
| EnterpriseBench: CoreCraft | Tool & computer use › Customer service & operations | 77.4 | 68.2 | −9.2 |
| Legal Research Bench | Professional › Legal | 56.3 | 47.1 | −9.1 |
| Artificial Analysis Strategy & Ops Index | Professional › General knowledge work | 63.8 | 54.8 | −8.9 |
| APEX-Agents 1.1 | Professional › General knowledge work | 88.9 | 80.4 | −8.5 |
| HANDBOOK.md Agents v1.1 | Tool & computer use › Customer service & operations | 52.3 | 44.0 | −8.3 |
| Finance Agent v2.1 | Professional › Finance, accounting & tax | 68.6 | 60.6 | −8.0 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 19.2 | 3.5 | −7.8 |
| DAYJOB: Tax | Professional › Finance, accounting & tax | 12.8 | 5.0 | −7.8 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 93.7 | 86.2 | −7.5 |
| Harvey LAB (Mercor run, undisclosed ~213-task cohort) | Professional › Legal | 26.6 | 19.5 | −7.1 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | -6.0 | 8.2 | +7.1 |
| Vals Legal Index | Professional › Legal | 78.4 | 72.7 | −5.7 |
| Excel Modeling Benchmark v1.1 | Professional › Finance, accounting & tax | 76.5 | 71.5 | −5.0 |
| GDP.pdf (November 2026 rescoring) | Professional › General knowledge work | 53.4 | 48.8 | −4.6 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1699 | 1634 | −3.5 |
| GPQA Diamond | Science › Multidisciplinary | 96.3 | 92.7 | −3.5 |
| OTIS Mock AIME 2024-2025 | Math | 100.0 | 98.1 | −1.9 |
| HLAB-Vals-1.1 | Professional › Legal | 19.6 | 18.3 | −1.2 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 1.0 | 1.2 | +0.1 |
Data: Artificial Analysis, Vals AI, Surge AI and 9 othersatlas.village.dev