DeepSeek V4.1 Flash 1017 vs GPT-6.2 Sol
GPT-6.2 Sol leads by 6.2 capability points (≈95% interval +3.5 to +8.9): a clear lead over DeepSeek V4.1 Flash 1017 · P(higher) ≈ 100%
Wins 53 of 68 shared benchmarks (15 lost)
Capability vs. relative cost
DeepSeek V4.1 Flash 1017GPT-6.2 SolClosed weightsOpen weights
Benchmark wins15 – 53 · 68 shared
Comparing matching setups only (32 benchmarks excluded) · suites counted once
Area← DeepSeek V4.1 Flash 1017GPT-6.2 Sol →
Data: Artificial Analysis, Vals AI, Surge AI and 11 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6.2 Sol minus DeepSeek V4.1 Flash 1017.
| ARC-AGI-3 (Semi-Private) | Games | 5.8 | 98.2 | +92.4 |
| CyberGym-E2E-AA (Artificial Analysis) | Software › Cybersecurity | 76.3 | 0.0 | −76.3 |
| Terminal-Bench-Science 0.1 | Science › Multidisciplinary | 12.9 | 67.6 | +54.8 |
| AA-PersonalityBench Organization index | Language & writing › Style & personality | 92.0 | 45.7 | −46.8 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 20.8 | 64.0 | +43.2 |
| RiemannBench (November 2026 rescoring) | Math | 38.4 | 80.0 | +41.6 |
| AA-PersonalityBench Informativeness index | Language & writing › Style & personality | 76.0 | 37.7 | −38.7 |
| AA-PersonalityBench Warmth index | Language & writing › Style & personality | 23.4 | 60.1 | +37.1 |
| AA-PersonalityBench Slop index | Language & writing › Style & personality | 48.9 | 15.4 | −33.9 |
| AA-PersonalityBench Transparency index | Language & writing › Style & personality | 88.8 | 56.9 | −32.3 |
| DAYJOB: Law (November 2026 release) | Professional › Legal | 2.8 | 32.7 | +29.9 |
| AA-PersonalityBench Imagination index | Language & writing › Style & personality | 31.3 | 60.1 | +29.0 |
| Terminal-Bench 4.0 | Software › Software engineering | 31.8 | 60.6 | +28.8 |
| Humanity's Sixth Sense | Perception & media › Spatial & visual reasoning | 30.1 | 56.8 | +26.7 |
| DeepsecBench-AA (Artificial Analysis) | Software › Cybersecurity | 24.9 | 51.6 | +26.6 |
| Terminal-Bench 4.0 (Artificial Analysis) | Software › Software engineering | 33.8 | 60.1 | +26.3 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 52.3 | 77.7 | +25.4 |
| Artificial Analysis Intelligence Index v5.0.1 | Composite indexes | 32.6 | 57.3 | +24.7 |
| Artificial Analysis Engineering Index | Science › Engineering | 37.6 | 61.2 | +23.6 |
| VIALS-AA (pass^3) | Science › Life sciences | 40.4 | 62.7 | +22.4 |
| EnterpriseBench: CoreCraft | Tool & computer use › Customer service & operations | 58.5 | 36.4 | −22.1 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 0.8 | 11.3 | +20.5 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 12.5 | 32.8 | +20.3 |
| DAYJOB: Accounting | Professional › Finance, accounting & tax | 7.6 | 27.6 | +20.0 |
| AA-Omniscience v1.1 Index | Knowledge & reasoning › General knowledge | 1.5 | 40.4 | +19.4 |
| Ramp SWE-Bench v2 (Bash harness) | Software › Software engineering | 53.4 | 72.4 | +19.0 |
| Artificial Analysis Economics Index | Composite indexes | 38.5 | 57.4 | +18.9 |
| DeepsecBench — Revalidated V1 (233 issues) | Software › Cybersecurity | 30.0 | 47.3 | +17.4 |
| Artificial Analysis Legal Index | Professional › Legal | 37.7 | 55.0 | +17.3 |
| Code Migration v1.1 (Vals Index v2.2 subset) | Software › Software engineering | 68.5 | 85.3 | +16.8 |
| AA-PersonalityBench Conviction index | Language & writing › Style & personality | 95.2 | 79.2 | −16.1 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 28.5 | 44.4 | +15.9 |
| VIALS-AA (average of 3) | Science › Life sciences | 53.2 | 68.9 | +15.7 |
| Artificial Analysis Finance & Accounting Index | Professional › Finance, accounting & tax | 42.2 | 57.2 | +15.1 |
| HealthBench Professional (Length-Adjusted) | Health & medicine | 51.7 | 65.9 | +14.2 |
| GDP.pdf (Artificial Analysis, All-pass) | Professional › General knowledge work | 17.2 | 30.0 | +12.8 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | -1.5 | 23.3 | +12.4 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 74.9 | 62.9 | −12.0 |
| Artificial Analysis Healthcare & Medical Index | Health & medicine | 41.3 | 53.3 | +12.0 |
| Bug Hunt Bench | Software › Software engineering | 40.2 | 52.1 | +11.9 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 6.0 | 17.2 | +11.2 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 26.7 | 37.2 | +10.6 |
| HANDBOOK.md Agents v1.1 | Tool & computer use › Customer service & operations | 24.0 | 34.5 | +10.5 |
| Artificial Analysis Strategy & Ops Index | Professional › General knowledge work | 49.2 | 59.4 | +10.2 |
| LiveBench 2.0 (2026-11-06 release) | Composite indexes | 70.3 | 80.3 | +10.0 |
| Legal Research Bench | Professional › Legal | 45.2 | 37.0 | −8.2 |
| VIALS-AA (pass@3) | Science › Life sciences | 66.5 | 74.5 | +8.1 |
| ARC-AGI-2 | Knowledge & reasoning › Abstract reasoning | 78.1 | 85.8 | +7.8 |
| CWE-Bench-AA (Artificial Analysis) | Software › Cybersecurity | 69.2 | 61.7 | −7.5 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1416 | 1541 | +6.8 |
| Agents' Last Exam v1.1 Score | Professional › General knowledge work | 66.6 | 73.0 | +6.4 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 2.3 | 13.4 | +5.5 |
| Vals Legal Index | Professional › Legal | 71.9 | 66.7 | −5.2 |
| Excel Modeling Benchmark v1.1 | Professional › Finance, accounting & tax | 69.2 | 74.3 | +5.1 |
| AA-PersonalityBench Expressiveness index | Language & writing › Style & personality | 50.5 | 55.3 | +4.8 |
| Vals Index v2.2 | Composite indexes | 61.3 | 66.1 | +4.8 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 2.2 | 10.9 | +4.3 |
| Vals Finance Index | Professional › Finance, accounting & tax | 60.9 | 56.7 | −4.2 |
| Finance Agent v2.1 | Professional › Finance, accounting & tax | 62.9 | 58.9 | −4.0 |
| GDP.xlsx | Professional › General knowledge work | 18.4 | 22.3 | +3.9 |
| DAYJOB: Tax | Professional › Finance, accounting & tax | 1.2 | 4.8 | +3.6 |
| ARC-AGI-1 Semi-Private | Knowledge & reasoning › Abstract reasoning | 94.5 | 97.5 | +3.0 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 68.5 | 71.3 | +2.8 |
| HLAB-Vals-1.1 | Professional › Legal | 14.2 | 12.1 | −2.1 |
| DualEntry Accounting AI Benchmark | Professional › Finance, accounting & tax | 82.2 | 84.2 | +2.0 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | 1.1 | 4.3 | +1.6 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | -0.1 | 1.2 | +0.6 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 87.7 | 87.6 | −0.1 |
Data: Artificial Analysis, Vals AI, Surge AI and 11 othersatlas.village.dev