GPT-5.6 Luna vs GPT-6 Luna
| Spec | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| Respan Index | 48.8 (#37) | 49.2 (#36) |
| Blended price per 1M | $0.225 | $0.1 |
| Context window | 1.1M | 1.1M |
| Input / output per 1M | $0.1 / $0.6 | $0.05 / $0.25 |
| Max output | 128K | 128K |
| Input | text, vision | text, vision |
| Released | 2026-07-09 | 2026-09-22 |
Respan Index by category
Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.
| Category | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| Coding | 27.8 GPT-5.6 Luna | 29.6 GPT-6 Luna |
| Agents | 56.9 GPT-5.6 Luna | 61.2 GPT-6 Luna |
| Reasoning | 62.6 GPT-5.6 Luna | 57.2 GPT-6 Luna |
| Knowledge & language | 38.7 GPT-5.6 Luna | 38.9 GPT-6 Luna |
| Human preference | 53.8 GPT-5.6 Luna | 55.2 GPT-6 Luna |
Benchmarks both report
GPT-5.6 Luna leads on 12 of the 23 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| Chess Puzzles (Epoch AI) | 43% | 36% |
| LMArena Elo | 1453 | 1441 |
| LMArena WebDev (LMArena) | 1519 | 1579 |
| LMArena Vision (LMArena) | 1257 | 1240 |
| LMArena Agent (LMArena) | -0.0081 | 0.0223 |
| APEX-Agents (Mercor) | 43% | 44.3% |
| FrontierMath Tier 4 v2 (Epoch AI) | 61% | 56.1% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 82.1% | 78.9% |
| Furniture Assembly (Epoch AI) | 42.5% | 44.2% |
| GPQA Diamond | 92.3% | 90.5% |
| LiveBench Agentic Coding (LiveBench) | 48.4% | 51.2% |
| LiveBench Coding (LiveBench) | 82.9% | 79% |
| LiveBench Data Analysis (LiveBench) | 78% | 73.4% |
| LiveBench Instruction Following (LiveBench) | 60.1% | 55.9% |
| LiveBench Language (LiveBench) | 72.6% | 73.8% |
| LiveBench Mathematics (LiveBench) | 87.2% | 89.1% |
| LiveBench Reasoning (LiveBench) | 85.6% | 81.8% |
| Mystery Game Puzzles (Epoch AI) | 21% | 7% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 98.3% | 98.9% |
| SAGE (Vals AI) | 44.2% | 48.1% |
| SimpleQA Verified | 41% | 41.4% |
| Terminal-Bench 4.0 (Vals AI) | 11.6% | 13.6% |
| DeepSWE v1.1 | 67.2% | 66.6% |
FAQ
Is GPT-5.6 Luna better than GPT-6 Luna?
GPT-6 Luna scores 49.2 on the Respan Index (#36), ahead of GPT-5.6 Luna at 48.8 (#37). GPT-5.6 Luna leads on 12 of the 23 benchmarks both report.
Which is cheaper, GPT-5.6 Luna or GPT-6 Luna?
GPT-5.6 Luna lists at $0.1 input and $0.6 output per 1M tokens; GPT-6 Luna at $0.05 and $0.25. GPT-6 Luna is 2.2x cheaper per token at list price.
Which has the larger context window?
Both accept 1.1M tokens of context.