Claude Opus 4.6 vs GPT-5.4
Seite an Seite. Jede Metrik. Jeder Benchmark.
| Typ | Claude Opus 4.6 | GPT-5.4 |
|---|---|---|
| Provider | ||
| Durchschn. Score | 57.5 | 59.0 |
| Eingabepreis | $5.00 | $2.50 |
| Ausgabepreis | $25.00 | $15.00 |
| Kontextfenster | 1.0M tokens (~500 books) | 1.1M tokens (~525 books) |
| Veroeffentlicht | 2026-02-04 | 2026-03-05 |
| Open Source | Proprietary | Proprietary |
Benchmark-Ergebnisse
13 benchmarks · Claude Opus 4.6: 6, GPT-5.4: 7
| Benchmark | Kategorie | Claude Opus 4.6 | GPT-5.4 |
|---|---|---|---|
| APEX-Agents | agentic | 31.7 | 35.9 |
| ARC-AGI | reasoning | 94.0 | 93.7 |
| ARC-AGI-2 | reasoning | 69.2 | 74.0 |
| Chatbot Arena Elo — Overall | arena | 1496.6 | 1465.8 |
| Chess Puzzles | knowledge | 17.0 | 44.0 |
| FrontierMath-2025-02-28-Private | math | 40.7 | 47.6 |
| FrontierMath-Tier-4-2025-07-01-Private | math | 22.9 | 27.1 |
| GPQA diamond | knowledge | 87.4 | 91.1 |
| OTIS Mock AIME 2024-2025 | math | 94.4 | 95.3 |
| PostTrainBench | knowledge | 23.2 | 20.2 |
| SimpleQA Verified | knowledge | 46.5 | 44.8 |
| SWE-Bench verified | coding | 78.7 | 76.9 |
| WeirdML | coding | 77.9 | 57.4 |