GPT-5.2
Reliable GPT generation for broad coding, writing, and tool-assisted product work
Shogo Agent Index
Supported#44 of 9246.4range 39.5–53.2 · 58% of weight measured
Agentic coding
49
57% measured
Long-horizon autonomy
59
80% measured
Knowledge-work agents
37
68% measured
Computer use
36
estimated
Security & context learning
46
50% measured
Agentic benchmarks
| Benchmark | Score | Normalised | Index weight | Harness / source |
|---|---|---|---|---|
| SWE-bench Verified | 73.8% | 8% | source | |
| Terminal-Bench 2.0 | 64.9% | 8% | Droid | |
| DeepSWE | est. | 6% | predicted 48 | |
| FrontierSWE | est. | 5% | predicted 15 | |
| CursorBench | est. | 4% | predicted 37 | |
| GSO-Bench | 27.5% | 4% | OpenHands | |
| METR Time Horizons | 75.3% | 10% | source | |
| Vending-Bench 2 | $3,591 | 6% | source | |
| PostTrainBench | est. | 4% | predicted 25 | |
| APEX-Agents | est. | 8% | predicted 44 | |
| GDPval | 49.7% | 7% | source | |
| Remote Labor Index | 2.5% | 5% | source | |
| DeepResearch Bench | 41.1% | 5% | source | |
| OSWorld 2.0 | est. | 6% | predicted 9 | |
| OSWorld | est. | 4% | predicted 75 | |
| Cybench | est. | 5% | predicted 74 | |
| CL-bench | 18.2% | 5% | source |
Other benchmarks
| Benchmark | Category | Score | Source |
|---|---|---|---|
| AlgoTune | Coding | 2.05× | source |
| ARC-AGI | Reasoning | 86.2% | source |
| ARC-AGI-2 | Reasoning | 52.9% | source |
| Chess Puzzles | Reasoning | 48.0% | source |
| DTBench | Knowledge | 90.9% | source |
| FrontierMath Tiers 1-3 | Math | 67.4% | source |
| FrontierMath Tier 4 | Math | 31.7% | source |
| GPQA Diamond | Reasoning | 91.4% | source |
| Humanity's Last Exam | Reasoning | 27.8% | source |
| LMCA | Knowledge | 43.9% | source |
| MirrorCode | Coding | 32.2% | source |
| OTIS Mock AIME | Math | 96.1% | source |
| ProofBench | Math | 15.0% | source |
| SimpleBench | Reasoning | 45.8% | source |
| SimpleQA Verified | Knowledge | 37.1% | source |
| VPCT | Multimodal | 84.0% | source |
| WebDev Arena | Coding | 1480 | source |
| WeirdML v3 | Coding | 5.0% | codex_cli |