GPT-5.6 Terra
Balanced GPT-5.6 model for capable, cost-efficient everyday work
Shogo Agent Index
Estimated#16 of 9258.8range 50.7–66.9 · 24% of weight measured
Agentic coding
66
29% measured
Long-horizon autonomy
69
30% measured
Knowledge-work agents
47
32% measured
Computer use
48
estimated
Security & context learning
56
estimated
Agentic benchmarks
| Benchmark | Score | Normalised | Index weight | Harness / source |
|---|---|---|---|---|
| SWE-bench Verified | est. | 8% | predicted 84 | |
| Terminal-Bench 2.0 | est. | 8% | predicted 87 | |
| DeepSWE | 69.6% | 6% | mini-swe-agent · $4.95/task | |
| FrontierSWE | est. | 5% | predicted 32 | |
| CursorBench | 41.3% | 4% | source · $5.14/task | |
| GSO-Bench | est. | 4% | predicted 50 | |
| METR Time Horizons | est. | 10% | predicted 81 | |
| Vending-Bench 2 | $7,343 | 6% | source | |
| PostTrainBench | est. | 4% | predicted 31 | |
| APEX-Agents | 58.2% | 8% | source | |
| GDPval | est. | 7% | predicted 58 | |
| Remote Labor Index | est. | 5% | predicted 8 | |
| DeepResearch Bench | est. | 5% | predicted 50 | |
| OSWorld 2.0 | est. | 6% | predicted 20 | |
| OSWorld | est. | 4% | predicted 88 | |
| Cybench | est. | 5% | predicted 86 | |
| CL-bench | est. | 5% | predicted 25 |
Other benchmarks
| Benchmark | Category | Score | Source |
|---|---|---|---|
| ARC-AGI | Reasoning | 96.5% | source |
| ARC-AGI-2 | Reasoning | 83.9% | source |
| BALROG | Long-horizon | 53.2% | source |
| Chess Puzzles | Reasoning | 47.0% | source |
| CritPt | Reasoning | 30.0% | source |
| DTBench | Knowledge | 93.3% | source |
| FrontierCode | Coding | 41.3% | codex |
| FrontierMath Tiers 1-3 | Math | 86.0% | source |
| FrontierMath Tier 4 | Math | 70.7% | source |
| GPQA Diamond | Reasoning | 93.3% | source |
| LMCA | Knowledge | 54.9% | source |
| MirrorCode | Coding | 41.1% | source |
| OTIS Mock AIME | Math | 99.7% | source |
| ProofBench | Math | 74.0% | source |
| SciCode | Coding | 55.0% | source |
| SimpleBench | Reasoning | 48.9% | source |
| SimpleQA Verified | Knowledge | 43.2% | source |
| WebDev Arena | Coding | 1521 | source |