GPT 5.6 Terra is GPT's balanced performance model, priced at $2.00/MTok input and $12.00/MTok output with a 1,050K context window.
Input $/MTok
$2.00
Output $/MTok
$12.00
Context window
1,050K
Cache read discount
90% off
| Rate | Input $/MTok | Output $/MTok |
|---|---|---|
| Standard | $2.00 | $12.00 |
| Cached read (90% off input) | $0.20 | - |
| Cached write (25% surcharge) | $2.50 | - |
| Batch (50% off) | $1.00 | $6.00 |
| Long context (≥272.001K tokens) | $4.00 | $18.00 |
All rates in USD per million tokens. Prices verified against provider documentation.
Confidence scores (0-100) and cross-provider rank for each task type. Rank 1 is best across all 20 models.
| Task | Confidence | Rank |
|---|---|---|
| Code Generation | 86/100 | #4 |
| Code Review | 84/100 | #4 |
| Summarization | 92/100 | #1 |
| Q&A | 92/100 | #1 |
| Extraction | 90/100 | #1 |
| Reasoning | 82/100 | #4 |
| Classification | 89/100 | #1 |
| Creative Writing | 80/100 | #5 |
Estimated cost per API call using default token counts for each workload.
| Use case | Input tokens | Output tokens | Cost / call |
|---|---|---|---|
| Code Generation | 2,000 | 4,000 | $0.0520 |
| Unit Test Generation | 3,000 | 5,000 | $0.0660 |
| SQL Generation | 1,500 | 1,000 | $0.0150 |
| API Generation | 3,000 | 6,000 | $0.0780 |
| Code Refactoring | 4,000 | 4,000 | $0.0560 |
| Code Review | 5,000 | 2,000 | $0.0340 |
| PR Review | 8,000 | 3,000 | $0.0520 |
| Security Review | 6,000 | 2,500 | $0.0420 |
| Bug Detection | 5,000 | 2,000 | $0.0340 |
| Document Summarization | 10,000 | 1,000 | $0.0320 |
| PDF Summarization | 15,000 | 1,500 | $0.0480 |
| Meeting Notes | 8,000 | 2,000 | $0.0400 |
Projected monthly spend at different call volumes, using code generation as a representative workload (2,000 input / 4,000 output tokens per call).
| Calls / month | Without caching | With caching (90% read discount) |
|---|---|---|
| 1K | $52 | $49.48 |
| 10K | $520 | $494.8 |
| 100K | $5,200 | $4,948 |
| 500K | $26,000 | $24,740 |
| 1M | $52,000 | $49,480 |
Caching estimate assumes 70% of input tokens are cache reads. Your actual cache hit rate will depend on prompt structure and reuse.
Same-tier models from other providers, sorted by input price.
| Model | Provider | Input $/MTok | Output $/MTok | Input savings |
|---|---|---|---|---|
| Grok 4.3 | Grok | $1.25 | $2.50 | 38% cheaper |
| Claude Sonnet 5 | Claude | $2.00 | $10.00 | Same price |
| Gemini 3.1 Pro | Gemini | $2.00 | $12.00 | Same price |
| Claude Sonnet 4.6 | Claude | $3.00 | $15.00 | 50% more |
Input decreased
$2.50 to $2.00/MTok
Output decreased
$15.00 to $12.00/MTok
See exactly how much you spend on GPT 5.6 Terra and where you can save. Two-line wrapper install, no API keys shared.