Cohere: Command A+ vs Sakana: Fugu Max
Side-by-side comparison of specs, pricing, benchmark scores, and task rankings. Updated 2026-09-27.
| Cohere: Command A+ | Sakana: Fugu Max | |
|---|---|---|
| Vendor | cohere | sakana |
| Quality Score | 100 | 100 |
| Benchmark Score | 17.4 | - |
| Input Price | $0.30/M | $2.00/M |
| Output Price | $1.50/M | $6.00/M |
| Context Window | 192,000 | 1,000,000 |
| Max Output | 64,000 | 128,000 |
| Tool Calling | ✓ | ✓ |
| Structured Output | ✓ | ✓ |
| Reasoning Mode | ✓ | ✓ |
| Vision | ✓ | ✓ |
| Audio | - | - |
| Benchmark Scores | ||
| ai_index | 21.7 | - |
The verdict
For a sample job of 1 million input tokens and 250,000 output tokens, Command A+ costs $0.68 and Fugu Max costs $3.50, so Command A+ is about 5.2 times cheaper for the same work.
Fugu Max reads up to 1,000,000 tokens at once, roughly 1,500 pages of text, against about 288 pages for Command A+.
Across the 20 tasks where they differ, Command A+ scores higher on 7 (Code Completion, Regex Writing, Frontend Component Design and others), Fugu Max on 13 (SQL Generation, Code Review, Code Refactoring and others).
Pick Command A+ if you need lower cost; pick Fugu Max if you need long documents.
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 | Cohere: Command A+ | Sakana: Fugu Max |
|---|---|---|
| SQL Generation | 132 | 133 |
| Code Review | 127 | 132 |
| Code Completion | 127 | 118 |
| Code Refactoring | 127 | 136 |
| Bug Fixing | 131 | 136 |
| Unit Test Generation | 122 | 124 |
| Code Documentation | 125 | 130 |
| Regex Writing | 121 | 118 |
| CI/CD Pipelines | 118 | 120 |
| Frontend Component Design | 124 | 122 |
| Data Analysis | 127 | 124 |
| CSV / Spreadsheet Cleanup | 126 | 133 |
| ETL Scripting | 122 | 128 |
| JSON Extraction | 131 | 122 |
| Bulk Data Labeling | 129 | 119 |
| OCR / Document Parsing | 127 | 131 |
| Table Extraction from PDFs | 127 | 131 |
| Long-Document Summarization | 128 | 137 |
| Short-Form Summarization | 124 | 114 |
| Blog Post Writing | 120 | 121 |
Scores reflect capability match + benchmark data + pricing for each task. Methodology →