Claude Opus 5.5 vs GPT-6.2 Astra
GPT-6.2 Astra leads by 0.9 capability points (≈95% interval −1.9 to +3.7): statistically tied · P(higher) ≈ 74%
Wins 41 of 70 shared benchmarks (26 lost, 3 tied)
Capability vs. relative cost
Claude Opus 5.5GPT-6.2 AstraClosed weightsOpen weights
Benchmark wins26 – 41 · 70 shared, 3 tied
Comparing matching setups only (63 benchmarks excluded) · suites counted once
Area← Claude Opus 5.5GPT-6.2 Astra →
Data: Artificial Analysis, Surge AI, Epoch AI and 18 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6.2 Astra minus Claude Opus 5.5.
| AA-PersonalityBench Informativeness index | Language & writing › Style & personality | 90.5 | 31.3 | −59.7 |
| AA-PersonalityBench Organization index | Language & writing › Style & personality | 79.1 | 21.8 | −57.9 |
| AA-PersonalityBench Transparency index | Language & writing › Style & personality | 90.5 | 40.9 | −50.1 |
| TxBench Preclinical Pharmacology V1 Clean Reruns | Science › Life sciences | 26.3 | 68.0 | +41.7 |
| SpireClimb | Games | 5.8 | 41.3 | +35.5 |
| DeepsecBench — Revalidated V1 (233 issues) | Software › Cybersecurity | 26.6 | 61.2 | +34.7 |
| AA-PersonalityBench Imagination index | Language & writing › Style & personality | 94.3 | 63.3 | −31.3 |
| EnterpriseBench: CoreCraft | Tool & computer use › Customer service & operations | 74.4 | 43.6 | −30.8 |
| MirrorCode (ML, +Private, 2L) | Software › Software engineering | 95.6 | 67.8 | −27.8 |
| AA-PersonalityBench Expressiveness index | Language & writing › Style & personality | 90.5 | 66.5 | −24.3 |
| Mystery Game Puzzles | Games | 71.0 | 95.0 | +24.0 |
| EBR-bench v4 (card ban, single agent) | Games | 71.4 | 92.4 | +21.0 |
| Humanity's Sixth Sense | Perception & media › Spatial & visual reasoning | 44.6 | 63.2 | +18.6 |
| DAYJOB: Law (November 2026 release) | Professional › Legal | 23.7 | 41.4 | +17.7 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 54.6 | 72.2 | +17.6 |
| SWE Atlas v1.1 - Overall | Software › Software engineering | 77.5 | 60.5 | −17.0 |
| AA-PersonalityBench Warmth index | Language & writing › Style & personality | 84.8 | 68.1 | −16.9 |
| Terminal-Bench-Science 0.1 | Science › Multidisciplinary | 61.9 | 78.6 | +16.7 |
| AA-3D v1.0 Elo | Software › Frontend & design | 1202 | 1318 | +15.7 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 48.3 | 34.4 | −13.9 |
| AA-PersonalityBench Slop index | Language & writing › Style & personality | 20.0 | 7.4 | −12.8 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 27.8 | 39.4 | +11.7 |
| SimpleBench | Knowledge & reasoning › General knowledge | 88.4 | 77.3 | −11.1 |
| Chess Puzzles (task version 2.0) | Games | 67.0 | 78.0 | +11.0 |
| Blueprint-Bench 2 | Perception & media › Spatial & visual reasoning | 0.5 | 0.6 | +10.2 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1800 | 1644 | −8.6 |
| sudo L7 (October 2026 rescoring) | Software › Software engineering | 46.1 | 38.3 | −7.8 |
| DAYJOB: Accounting | Professional › Finance, accounting & tax | 34.4 | 42.0 | +7.6 |
| FrontierMath Erdős | Math | 2.9 | 10.3 | +7.4 |
| HLAB-Vals-1.1 | Professional › Legal | 9.2 | 16.3 | +7.1 |
| GDPval v2 (OpenAI internal) | Professional › General knowledge work | 43.5 | 50.4 | +6.9 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 18.6 | 15.2 | −6.6 |
| HANDBOOK.md Agents v1.1 | Tool & computer use › Customer service & operations | 52.3 | 45.8 | −6.5 |
| SimpleQA Verified (Epoch anti-abstention) | Knowledge & reasoning › General knowledge | 72.2 | 78.5 | +6.3 |
| AA-AnalystAgent (pass^5) | Professional › General knowledge work | 56.3 | 62.5 | +6.3 |
| Furniture Assembly Benchmark | Perception & media › Spatial & visual reasoning | 83.3 | 89.2 | +5.8 |
| BigFinanceBench private holdout (final-answer accuracy) | Professional › Finance, accounting & tax | 48.2 | 54.0 | +5.8 |
| ComplexConstraints (November 2026 rescoring) | Language & writing › Instruction following | 64.4 | 69.6 | +5.2 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 43.8 | 49.0 | +5.2 |
| FrontierMath Tier 4 (v2) - Private | Math | 95.0 | 100.0 | +5.0 |
| Terminal-Bench 4.0 (Artificial Analysis) | Software › Software engineering | 59.6 | 64.1 | +4.5 |
| Professional Reasoning Benchmark - Legal | Professional › Legal | 47.6 | 43.1 | −4.5 |
| DualEntry Accounting AI Benchmark | Professional › Finance, accounting & tax | 82.2 | 78.2 | −4.0 |
| RiemannBench (November 2026 rescoring) | Math | 80.0 | 76.0 | −4.0 |
| Ramp SWE-Bench v2 (Bash harness) | Software › Software engineering | 71.2 | 75.0 | +3.8 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 19.2 | 12.4 | −3.4 |
| Epoch Capabilities Index (ECI) | Composite indexes | 169 | 173 | +3.3 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 42.5 | 45.8 | +3.3 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | 14.0 | 7.7 | −3.2 |
| MMMU Pro | Knowledge & reasoning › General knowledge | 86.6 | 89.5 | +2.9 |
| ARC-AGI-2 | Knowledge & reasoning › Abstract reasoning | 93.3 | 95.8 | +2.5 |
| Dual-Arm Sim (task completion) | Tool & computer use › Robotics | 0.0 | 2.5 | +2.5 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 81.7 | 83.7 | +2.0 |
| TaxCalcBench TY25 | Professional › Finance, accounting & tax | 42.0 | 44.0 | +2.0 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 18.0 | 14.0 | −2.0 |
| AA-PersonalityBench Conviction index | Language & writing › Style & personality | 69.5 | 71.3 | +1.7 |
| LiveBench 2.0 (2026-11-06 release) | Composite indexes | 83.4 | 84.9 | +1.4 |
| FrontierMath Tiers 1-3 (v2) - Private | Math | 91.2 | 92.6 | +1.4 |
| Gray Swan IPI Attack Success (k=15) | Safety & alignment | 1.0 | 2.4 | −1.4 |
| Chartography (November 2026 rescoring) | Perception & media › Documents & charts | 89.6 | 90.7 | +1.1 |
| LLM Debate Benchmark | Language & writing › Language & wordplay | 1680 | 1673 | −1.0 |
| GDP.pdf (Artificial Analysis, All-pass) | Professional › General knowledge work | 28.8 | 29.8 | +1.0 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 38.6 | 37.0 | −0.8 |
| GDP.xlsx | Professional › General knowledge work | 30.3 | 29.7 | −0.6 |
| GDP.pdf (November 2026 rescoring) | Professional › General knowledge work | 48.8 | 49.2 | +0.4 |
| Excel Modeling Benchmark v1.1 | Professional › Finance, accounting & tax | 75.8 | 76.0 | +0.2 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 1.0 | 1.2 | +0.1 |
| ARC-AGI-1 Semi-Private | Knowledge & reasoning › Abstract reasoning | 98.5 | 98.5 | 0.0 |
| DAYJOB: Tax | Professional › Finance, accounting & tax | 9.2 | 9.2 | 0.0 |
| OTIS Mock AIME 2024-2025 | Math | 100.0 | 100.0 | 0.0 |
Data: Artificial Analysis, Surge AI, Epoch AI and 18 othersatlas.village.dev