head-to-head

Mistral: Mistral Medium 3.5 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	Mistral: Mistral Medium 3.5	Anthropic: Claude Sonnet 4.6
SQL Generation	133	154
Code Review	130	161
Code Completion	115	120
Code Refactoring	131	162
Bug Fixing	134	166
Unit Test Generation	124	144
Code Documentation	126	142
Regex Writing	120	129
CI/CD Pipelines	120	138
Frontend Component Design	124	137
Data Analysis	127	147
CSV / Spreadsheet Cleanup	128	145
ETL Scripting	125	150
JSON Extraction	122	125
Bulk Data Labeling	119	118
OCR / Document Parsing	129	140
Table Extraction from PDFs	129	140
Long-Document Summarization	132	161
Short-Form Summarization	115	118
Blog Post Writing	121	137

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