Claude Sonnet 4.5 (batch) vs GPT-5.6 Terra
Side-by-side API pricing and Artificial Analysis performance for Claude Sonnet 4.5 (batch) (Anthropic) and GPT-5.6 Terra (OpenAI). Green cells mark the better value in each row.
Prices updated Sep 16, 2026, 1:42 AM UTC · refreshed hourly
| Claude Sonnet 4.5 (batch) Anthropic | GPT-5.6 Terra OpenAI | |
|---|---|---|
| Pricing | ||
| Input $/1M | $1.50 | $2.00 |
| Output $/1M | $7.50 | $12.00 |
| Blended $/1M (3:1) | $3.00 | $4.50 |
| RAG example (30K in / 2K out) | $0.06 | $0.084 |
| Quality & speed | ||
| Intelligence Index | 21.2 | 42.3 |
| Coding Index | 52.1 | 76.7 |
| Agentic Index | 17.5 | 43.7 |
| Output speed (tok/s) | — | 119 |
| Time to first token | — | 102.93s |
| Specs | ||
| Context window | 1M | 1.05M |
| Input modalities | text, image, file | file, image, text |
FAQ
Which is cheaper, Claude Sonnet 4.5 (batch) or GPT-5.6 Terra?
Claude Sonnet 4.5 (batch) has the lower blended API price at $3.00 per 1M tokens (3:1 input:output mix), versus $4.50 for the other model. Prices are live from OpenRouter and refresh about hourly.
Which is better for RAG workloads?
For a typical RAG request (30K input / 2K output tokens), Claude Sonnet 4.5 (batch) costs about $0.06 per request versus $0.084. Also compare context windows: Claude Sonnet 4.5 (batch) offers 1M and GPT-5.6 Terra offers 1.05M.
Which scores higher on Artificial Analysis benchmarks?
GPT-5.6 Terra leads on the Artificial Analysis Intelligence Index (42.3 vs 21.2). Check coding and agentic indices on this page for workload-specific tradeoffs.