We ran both ChatGPT-5 (OpenAI) and Claude-4 (Anthropic) through 50 tasks across 5 categories: coding (10), creative writing (10), data analysis (10), reasoning (10), and instruction following (10). Each task was scored 0-10 by two independent evaluators.
| Category | ChatGPT-5 | Claude-4 |
|---|---|---|
| Coding | 8.7 | 8.1 |
| Creative Writing | 7.8 | 9.1 |
| Data Analysis | 8.9 | 8.3 |
| Reasoning | 8.5 | 8.8 |
| Instructions | 8.2 | 9.0 |
Coding: ChatGPT-5 pulls ahead with better debugging and multi-file refactoring. It catches edge cases Claude misses.
Creative Writing: Claude-4 produces more natural, less formulaic prose. Its tone adaptation is noticeably better.
Verdict: For developers, pick ChatGPT. For content creators, Claude. For general use, the gap is narrow enough that your preference for the ecosystem (OpenAI vs Anthropic) matters more than raw scores.