GPT-5.6 Luna
Cost-efficient GPT-5.6 model for fast, high-volume workloads
Shogo Agent Index
Estimated#34 of 9251.4range 42.8–59.9 · 24% of weight measured
Agentic coding
59
29% measured
Long-horizon autonomy
61
30% measured
Knowledge-work agents
39
32% measured
Computer use
40
estimated
Security & context learning
51
estimated
Agentic benchmarks
| Benchmark | Score | Normalised | Index weight | Harness / source |
|---|---|---|---|---|
| SWE-bench Verified | est. | 8% | predicted 80 | |
| Terminal-Bench 2.0 | est. | 8% | predicted 78 | |
| DeepSWE | 67.2% | 6% | mini-swe-agent · $3.03/task | |
| FrontierSWE | est. | 5% | predicted 20 | |
| CursorBench | 35.9% | 4% | source · $1.03/task | |
| GSO-Bench | est. | 4% | predicted 35 | |
| METR Time Horizons | est. | 10% | predicted 77 | |
| Vending-Bench 2 | $4,095 | 6% | source | |
| PostTrainBench | est. | 4% | predicted 27 | |
| APEX-Agents | 43.0% | 8% | source | |
| GDPval | est. | 7% | predicted 51 | |
| Remote Labor Index | est. | 5% | predicted 5 | |
| DeepResearch Bench | est. | 5% | predicted 49 | |
| OSWorld 2.0 | est. | 6% | predicted 12 | |
| OSWorld | est. | 4% | predicted 81 | |
| Cybench | est. | 5% | predicted 79 | |
| CL-bench | est. | 5% | predicted 23 |
Other benchmarks
| Benchmark | Category | Score | Source |
|---|---|---|---|
| ARC-AGI | Reasoning | 88.0% | source |
| ARC-AGI-2 | Reasoning | 59.5% | source |
| BALROG | Long-horizon | 45.6% | source |
| Chess Puzzles | Reasoning | 43.0% | source |
| CritPt | Reasoning | 20.6% | source |
| DTBench | Knowledge | 89.1% | source |
| FrontierCode | Coding | 39.8% | codex |
| FrontierMath Tiers 1-3 | Math | 82.1% | source |
| FrontierMath Tier 4 | Math | 61.0% | source |
| GPQA Diamond | Reasoning | 91.6% | source |
| LMCA | Knowledge | 48.5% | source |
| MirrorCode | Coding | 21.1% | source |
| OTIS Mock AIME | Math | 98.3% | source |
| ProofBench | Math | 60.0% | source |
| SciCode | Coding | 53.6% | source |
| SimpleBench | Reasoning | 46.8% | source |
| SimpleQA Verified | Knowledge | 41.0% | source |
| WebDev Arena | Coding | 1519 | source |
| WeirdML v3 | Coding | 5.2% | codex_cli |
API hosts
Live prices and uptime for every host serving GPT-5.6 Luna via OpenRouter.
| Host | Input | Output | Cache read | Agent $/M | Uptime 24h | Quant |
|---|---|---|---|---|---|---|
| OpenAI | $0.10 | $0.60 | $0.010 | $0.050 | 100.0% | — |
| Amazon Bedrock | $0.20 | $1.20 | $0.020 | $0.10 | 100.0% | — |
| Azure | $0.20 | $1.20 | $0.020 | $0.10 | 100.0% | — |
| OpenAI | $0.20 | $1.20 | $0.020 | $0.10 | 100.0% | — |
| OpenAI | $0.40 | $2.40 | $0.040 | $0.20 | 100.0% | — |