Thinking Machines: Inkling Small vs Google: Gemini 3.1 Flash Lite
Side-by-side comparison of specs, pricing, benchmark scores, and task rankings. Updated 2026-11-07.
| Thinking Machines: Inkling Small | Google: Gemini 3.1 Flash Lite | |
|---|---|---|
| Vendor | thinkingmachines | |
| Quality Score | 100 | 100 |
| Input Price | $0.45/M | $0.25/M |
| Output Price | $1.20/M | $1.50/M |
| Context Window | 524,288 | 1,048,576 |
| Max Output | 262,144 | 65,536 |
| Tool Calling | ✓ | ✓ |
| Structured Output | - | ✓ |
| Reasoning Mode | ✓ | ✓ |
| Vision | ✓ | ✓ |
| Audio | ✓ | ✓ |
The verdict
For a sample job of 1 million input tokens and 250,000 output tokens, Gemini 3.1 Flash Lite costs $0.62 and Inkling Small costs $0.75, so Gemini 3.1 Flash Lite is about 1.2 times cheaper for the same work.
Gemini 3.1 Flash Lite reads up to 1,048,576 tokens at once, roughly 1,573 pages of text, against about 786 pages for Inkling Small.
Only Gemini 3.1 Flash Lite offers structured output and video input.
Across the 13 tasks where they differ, Gemini 3.1 Flash Lite on 13 (SQL Generation, Code Review, Code Refactoring and others).
Inkling Small was released in July 2026 and Gemini 3.1 Flash Lite in May 2026.
Pick Gemini 3.1 Flash Lite if you need lower cost, long documents, structured output, video input or the stronger all-round task scores.
Costs use current list prices per million tokens; page counts assume about 0.75 words per token and 500 words per page.
Who wins by task?
| Task | Thinking Machines: Inkling Small | Google: Gemini 3.1 Flash Lite |
|---|---|---|
| SQL Generation | 125 | 133 |
| Code Review | 128 | 132 |
| Code Refactoring | 132 | 136 |
| Unit Test Generation | 116 | 124 |
| CI/CD Pipelines | 116 | 120 |
| Frontend Component Design | 118 | 122 |
| Data Analysis | 120 | 124 |
| CSV / Spreadsheet Cleanup | 121 | 133 |
| ETL Scripting | 124 | 128 |
| JSON Extraction | 119 | 131 |
| Bulk Data Labeling | 125 | 129 |
| OCR / Document Parsing | 123 | 131 |
| Table Extraction from PDFs | 123 | 131 |
Scores reflect capability match + benchmark data + pricing for each task. Methodology →