OpenAI · GPT-5.6
GPT-5.6 Luna
Small GPT-5.6 model still offered via API; superseded by the cheaper GPT-6 Luna.
- AA Intelligence
- 37.3#28 of 142
- Input / output per 1M
- $0.20/ $1.20
- Output speed
- 126tok/s
- Context window
- 1.05M tokens
Composite indexes
- Artificial Analysis Intelligence Index#28 of 14237
Leader: Claude Opus 5.5 at 58
Coding
- Terminal-Bench 4.0#30 of 8811.6%
Leader: Claude Mythos 5.1 at 60.9%
- Terminal-Bench 2.x#21 of 10380.9%
Leader: Claude Fable 5.1 at 91.4%
- SciCode#28 of 9453.6%
Leader: Claude Opus 5.5 at 66.9%
Agents and tool use
- GDPval-AA#23 of 1031443
Leader: Claude Opus 5.5 at 1846
Reasoning and knowledge
- Humanity’s Last Exam#36 of 14739.5%
Leader: Claude Opus 5.5 at 61.4%
- GPQA Diamond#31 of 14291.1%
Leader: GPT-6 Astra at 96.1%
- CritPt#15 of 13920.6%
Leader: GPT-5.6 Sol at 32.3%
- ARC-AGI-2#15 of 2559.5%
Leader: GPT-6 Astra at 95%
- ARC-AGI-3#8 of 90.2%
Leader: GPT-6 Astra at 62.7%
Vision and long context
- AA Long Context Reasoning#9 of 13983.7%
Leader: Kimi K3 at 88.7%
- MMMU-Pro#23 of 7278.6%
Leader: Claude Opus 5.5 at 87.7%
Human preference
- LMArena Text#33 of 541452
Leader: Claude Fable 5 at 1506
- LMArena WebDev#33 of 541519
Leader: Claude Opus 5.5 at 1818
Other published results
| Result | Score | Reported by | Source |
|---|---|---|---|
| AA-Omniscience Index · max effort | -10.3 index (-100 to 100) | independent | Artificial Analysis: GPT-5.6 Luna (max) ↗ |
| AutomationBench-AA (partial score) · max effort | 50.2% | independent | Artificial Analysis: GPT-5.6 Luna (max) ↗ |
| Terminal-Bench Science (AA) · max effort | 3.3% | independent | Artificial Analysis: GPT-5.6 Luna (max) ↗ |
| AA Coding Agent Index v1.5 · Codex + GPT-5.6 Luna (max); DeepSWE 66.4 / SWE-Atlas-QnA 48.7 / TB4 14.6 | 43.2 index | independent | Artificial Analysis Coding Agents ↗ |