GPT-6.2 Sol vs Grok 4.6
GPT-6.2 Sol leads by 7.2 capability points (≈95% interval +4.8 to +9.6): a clear lead over Grok 4.6 · P(higher) ≈ 100%
Wins 37 of 49 shared benchmarks (11 lost, 1 tied)
Capability vs. relative cost
GPT-6.2 SolGrok 4.6Closed weightsOpen weights
Benchmark wins37 – 11 · 49 shared, 1 tied
Comparing matching setups only (6 benchmarks excluded) · suites counted once
Area← GPT-6.2 SolGrok 4.6 →
Data: Artificial Analysis, Vals AI, Arena and 11 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Grok 4.6 minus GPT-6.2 Sol.
| Terminal-Bench-Science 0.1 | Science › Multidisciplinary | 67.6 | 6.2 | −61.4 |
| Chartography | Perception & media › Documents & charts | 75.1 | 20.5 | −54.6 |
| Gray Swan IPI Attack Success (k=15) | Safety & alignment | 2.3 | 51.8 | −49.5 |
| Terminal-Bench Science 0.1 (Vals) | Science › Multidisciplinary | 58.6 | 11.4 | −47.1 |
| Terminal-Bench 4.0 | Software › Software engineering | 60.6 | 17.2 | −43.4 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 64.0 | 21.5 | −42.5 |
| AA-PersonalityBench Imagination index | Language & writing › Style & personality | 60.1 | 20.0 | −40.4 |
| ARC-AGI-3 (Semi-Private) | Games | 40.1 | 2.1 | −38.0 |
| Terminal-Bench 4.0 (Artificial Analysis) | Software › Software engineering | 58.6 | 21.2 | −37.4 |
| AA-PersonalityBench Organization index | Language & writing › Style & personality | 45.7 | 14.3 | −31.7 |
| DeepsecBench — Revalidated V1 (233 issues) | Software › Cybersecurity | 47.3 | 16.1 | −31.3 |
| EnterpriseBench: CoreCraft | Tool & computer use › Customer service & operations | 36.4 | 65.6 | +29.2 |
| AA-PersonalityBench Warmth index | Language & writing › Style & personality | 60.1 | 31.5 | −28.9 |
| Artificial Analysis Intelligence Index v5.0.1 | Composite indexes | 57.3 | 31.0 | −26.2 |
| EnterpriseOps-Gym-AA | Tool & computer use › Customer service & operations | 23.4 | 48.3 | +25.0 |
| Blueprint-Bench 2 | Perception & media › Spatial & visual reasoning | 0.6 | 0.3 | −23.8 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 11.3 | -0.6 | −23.2 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 42.8 | 19.8 | −23.0 |
| AA-PersonalityBench Slop index | Language & writing › Style & personality | 15.4 | 35.3 | +20.1 |
| Vals Science Index | Science › Multidisciplinary | 57.8 | 38.1 | −19.7 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 37.2 | 18.3 | −18.9 |
| ARC-AGI-2 | Knowledge & reasoning › Abstract reasoning | 85.8 | 67.1 | −18.7 |
| LiveBench 2.0 (2026-11-06 release) | Composite indexes | 80.3 | 63.5 | −16.8 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 78.3 | 62.3 | −16.0 |
| AA-PersonalityBench Transparency index | Language & writing › Style & personality | 56.9 | 71.4 | +14.7 |
| HealthBench Professional (Length-Adjusted) | Health & medicine | 65.9 | 51.7 | −14.2 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 23.3 | -3.5 | −13.4 |
| AA-PersonalityBench Conviction index | Language & writing › Style & personality | 79.2 | 92.4 | +13.3 |
| GDP.pdf (Artificial Analysis, All-pass) | Professional › General knowledge work | 30.0 | 17.8 | −12.2 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 32.8 | 21.0 | −11.8 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 87.6 | 76.2 | −11.4 |
| AA-AnalystAgent (pass^5) | Professional › General knowledge work | 52.5 | 41.3 | −11.2 |
| Legal Research Bench | Professional › Legal | 37.0 | 48.1 | +11.1 |
| ARC-AGI-1 Semi-Private | Knowledge & reasoning › Abstract reasoning | 97.5 | 87.5 | −10.0 |
| AA-PersonalityBench Expressiveness index | Language & writing › Style & personality | 55.3 | 63.8 | +8.6 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 62.9 | 71.4 | +8.5 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 13.4 | -2.5 | −7.9 |
| AA-Omniscience v1.1 Index | Knowledge & reasoning › General knowledge | 40.4 | 24.9 | −7.8 |
| AA-PersonalityBench Informativeness index | Language & writing › Style & personality | 37.7 | 44.8 | +7.1 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 10.9 | -2.5 | −6.7 |
| MysteryMechanism | Math | 36.9 | 30.6 | −6.3 |
| SimpleBench | Knowledge & reasoning › General knowledge | 69.9 | 75.9 | +6.0 |
| BioMysteryBench (Vals Terminus 2) | Science › Life sciences | 77.8 | 72.2 | −5.6 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 71.3 | 67.0 | −4.3 |
| TxBench Preclinical Pharmacology V1 Clean Reruns | Science › Life sciences | 65.3 | 61.7 | −3.7 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | 4.3 | 9.1 | +2.4 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 17.2 | 14.8 | −2.4 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1580 | 1540 | −2.2 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 1.2 | 1.2 | 0.0 |
Data: Artificial Analysis, Vals AI, Arena and 11 othersatlas.village.dev