Side-by-side technical and economic comparison between Anthropic's Claude Opus 5 and OpenAI's GPT-5.6 Luna.
| Metric | Claude Opus 5 | GPT-5.6 Luna | Advantage |
|---|---|---|---|
| Provider Organization | Anthropic | OpenAI | — |
| Standard Input / 1M | $5.00 | $0.20 | GPT-5.6 Luna (96% lower) |
| Cached Input / 1M | $0.50 | $0.02 | GPT-5.6 Luna lower |
| Output / 1M | $25.00 | $1.20 | GPT-5.6 Luna lower |
| Context Window | 1M | 1.05M | GPT-5.6 Luna (1.05M) |
| Max Generation Tokens | 16.4k | 16.4k | Claude Opus 5 |
| Tokenizer Family | Anthropic Claude BPE (~65k vocabulary) | OpenAI o200k_base (200k vocabulary) | — |
| Workload (per month) | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| Chatbot10k requests · 500 in / 200 out | $75.00 | $3.40 |
| RAG app100k requests · 8,000 in / 1,200 out | $7,000.00 | $304.00 |
| Coding agent100k requests · 12,000 in / 2,000 out | $11,000.00 | $480.00 |
Standard rates, no caching or batch discounts. Model your own numbers below.
Opt for Claude Opus 5 when your engineering requirements prioritize Anthropic's ecosystem, specific tokenizer efficiencies (Calibrated Claude Tokenizer (±3%)), or when your expected prompt-to-completion ratios favor its $5/M input rate.
Opt for GPT-5.6 Luna when looking for OpenAI's tooling integration, specific context window depth (1.05M tokens), or when output generation volume favors its $1.2/M rate.
At a typical 3:1 input-to-output token mix, GPT-5.6 Luna is about 96% cheaper than Claude Opus 5. Claude Opus 5 costs $5 per 1M input and $25 per 1M output tokens; GPT-5.6 Luna costs $0.2 and $1.2.
GPT-5.6 Luna has the larger context window: 1.05M tokens vs 1M.
Claude Opus 5: yes, cached input costs $0.5/1M. GPT-5.6 Luna: yes, cached input costs $0.02/1M.