head-to-head

Thinking Machines: Inkling Small vs StepFun: Step 3.7 Flash

Side-by-side comparison of specs, pricing, benchmark scores, and task rankings. Updated 2026-08-13.

	Thinking Machines: Inkling Small	StepFun: Step 3.7 Flash
Vendor	thinkingmachines	stepfun
Quality Score	100	100
Benchmark Score	69.1	50.1
Input Price	$0.45/M	$0.20/M
Output Price	$1.20/M	$1.15/M
Context Window	524,288	262,144
Max Output	262,144	256,000
Tool Calling	✓	✓
Structured Output	✓	✓
Reasoning Mode	✓	✓
Vision	✓	✓
Audio	✓	-
Benchmark Scores
ai_index	67.9	51.0
ai_index_agentic	52.6	35.9
ai_index_coding	87.4	65.3

Who wins by task?

Task	Thinking Machines: Inkling Small	StepFun: Step 3.7 Flash
SQL Generation	164	153
Code Review	159	146
Code Completion	133	130
Code Refactoring	158	144
Bug Fixing	170	155
Unit Test Generation	148	139
Code Documentation	141	133
Regex Writing	133	129
CI/CD Pipelines	139	131
Frontend Component Design	141	136
Data Analysis	160	150
CSV / Spreadsheet Cleanup	151	141
ETL Scripting	149	138
JSON Extraction	146	142
Bulk Data Labeling	134	133
OCR / Document Parsing	144	138
Table Extraction from PDFs	144	138
Long-Document Summarization	155	142
Short-Form Summarization	130	128
Blog Post Writing	136	130

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