head-to-head

OpenAI: GPT Mini Latest vs Anthropic: Claude Sonnet 4.6

Side-by-side comparison of specs, pricing, benchmark scores, and task rankings. Updated 2026-09-23.

Who wins by task?

Task	OpenAI: GPT Mini Latest	Anthropic: Claude Sonnet 4.6
SQL Generation	133	154
Code Review	132	161
Code Completion	130	120
Code Refactoring	136	162
Bug Fixing	136	166
Unit Test Generation	124	144
Code Documentation	130	142
Regex Writing	119	129
CI/CD Pipelines	120	138
Frontend Component Design	122	137
Data Analysis	124	147
CSV / Spreadsheet Cleanup	133	145
ETL Scripting	128	150
JSON Extraction	130	125
Bulk Data Labeling	128	118
OCR / Document Parsing	131	140
Table Extraction from PDFs	131	140
Long-Document Summarization	137	161
Short-Form Summarization	123	118
Blog Post Writing	121	137

Scores reflect capability match + benchmark data + pricing for each task. Methodology →