GPT-5.6 Luna vs GPT-5.6 Terra.
Verdict by category, tale of the tape, shared benchmarks only, and cost at three workloads — all from the tracked data.
Verdict by category
| Category | GPT-5.6 Luna | GPT-5.6 Terra | Winner |
|---|---|---|---|
| League Score | 70.4 | 71.8 | GPT-5.6 Terra |
| Coding (SWE-bench %) | — | — | — |
| Reasoning (GPQA %) | 92.3 | 92.9 | GPT-5.6 Terra |
| Context (tokens) | 1050000 | 1050000 | Draw |
| Speed (tok/s) | — | — | — |
| Price (blended $/1M) | $0.45 | $4.50 | GPT-5.6 Luna (cheaper) |
Tale of the tape
| Spec | GPT-5.6 Luna | GPT-5.6 Terra |
|---|---|---|
| Maker | OpenAI | OpenAI |
| Released | Jul 9, 2026 | Jul 9, 2026 |
| League Score | 70 | 72 |
| Context | 1,050,000 tokens | 1,050,000 tokens |
| Max output | 128,000 | 128,000 |
| $/1M input | $0.20 | $2.00 |
| $/1M output | $1.20 | $12.00 |
| Weights | Closed | Closed |
| Speed | — | — |
Head-to-head benchmarks (shared only)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Terra |
|---|---|---|
| LiveBench overall | 66.1 | 67.2 |
| GPQA Diamond | 92.3 | 92.9 |
Cost at three workloads
| Workload | GPT-5.6 Luna | GPT-5.6 Terra |
|---|---|---|
| Light — 1M in / 250K out | $0.50 | $5.00 |
| Standard — 10M in / 2.5M out | $5.00 | $50.00 |
| Long-context — 50M in / 5M out | $16.00 | $160.00 |
The pick
Pick GPT-5.6 Luna if you want the lower running cost.
Pick GPT-5.6 Terra if you want the higher overall League Score.
All figures from the model database — sources and dates on each model page. Cost your own workload or compare a different pair.
Frequently asked questions
Which is better overall, GPT-5.6 Luna or GPT-5.6 Terra?
GPT-5.6 Terra leads on overall League Score, 72 to 70 — a gap of 1 points on a composite of sourced benchmarks. That is overall form, not every category: check the per-category verdicts before choosing.
Which is cheaper, GPT-5.6 Luna or GPT-5.6 Terra?
GPT-5.6 Luna is cheaper: $0.45 vs $4.50 blended per 1M tokens (3:1 input/output mix) — about 10.0x less per token.
Which has the bigger context window, GPT-5.6 Luna or GPT-5.6 Terra?
Neither — both take 1.1M of context.
Which is better at reasoning, GPT-5.6 Luna or GPT-5.6 Terra?
GPT-5.6 Terra leads on GPQA Diamond, 92.9% vs 92.3% on graduate-level science questions.
Are GPT-5.6 Luna and GPT-5.6 Terra open source?
GPT-5.6 Luna is closed weights — API access only; GPT-5.6 Terra is closed weights — API access only.