OpenAI: GPT-5.6 Terra Pro vs OpenAI: GPT-5.6 Sol
Side-by-side comparison of specs, pricing, benchmark scores, and task rankings. Updated 2026-10-02.
| OpenAI: GPT-5.6 Terra Pro | OpenAI: GPT-5.6 Sol | |
|---|---|---|
| Vendor | openai | openai |
| Quality Score | 100 | 100 |
| Benchmark Score | - | 80.2 |
| Input Price | $2.00/M | $2.00/M |
| Output Price | $12.00/M | $10.00/M |
| Context Window | 1,050,000 | 1,050,000 |
| Max Output | 128,000 | 128,000 |
| Tool Calling | ✓ | ✓ |
| Structured Output | ✓ | ✓ |
| Reasoning Mode | ✓ | ✓ |
| Vision | ✓ | ✓ |
| Audio | - | - |
| Benchmark Scores | ||
| ai_index | - | 77.5 |
The verdict
For a sample job of 1 million input tokens and 250,000 output tokens, GPT-5.6 Sol costs $4.50 and GPT-5.6 Terra Pro costs $5.00, so GPT-5.6 Sol is slightly cheaper for the same work.
Both read up to 1,050,000 tokens at once, roughly 1,575 pages of text.
Across the 20 tasks where they differ, GPT-5.6 Sol on 20 (SQL Generation, Code Review, Code Completion and others).
Pick GPT-5.6 Sol if you need lower cost 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 | OpenAI: GPT-5.6 Terra Pro | OpenAI: GPT-5.6 Sol |
|---|---|---|
| SQL Generation | 132 | 142 |
| Code Review | 132 | 146 |
| Code Completion | 117 | 119 |
| Code Refactoring | 136 | 148 |
| Bug Fixing | 136 | 150 |
| Unit Test Generation | 124 | 133 |
| Code Documentation | 129 | 136 |
| Regex Writing | 117 | 125 |
| CI/CD Pipelines | 120 | 129 |
| Frontend Component Design | 122 | 131 |
| Data Analysis | 124 | 135 |
| CSV / Spreadsheet Cleanup | 132 | 135 |
| ETL Scripting | 128 | 138 |
| JSON Extraction | 121 | 121 |
| Bulk Data Labeling | 117 | 118 |
| OCR / Document Parsing | 131 | 134 |
| Table Extraction from PDFs | 131 | 134 |
| Long-Document Summarization | 136 | 149 |
| Short-Form Summarization | 113 | 117 |
| Blog Post Writing | 120 | 130 |
Scores reflect capability match + benchmark data + pricing for each task. Methodology →