GPT-5
Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows
Shogo Agent Index
Supported#60 of 9237.8range 27.8–47.8 · 47% of weight measured
Agentic coding
40
57% measured
Long-horizon autonomy
47
50% measured
Knowledge-work agents
31
68% measured
Computer use
27
estimated
Security & context learning
40
estimated
Agentic benchmarks
| Benchmark | Score | Normalised | Index weight | Harness / source |
|---|---|---|---|---|
| SWE-bench Verified | 73.6% | 8% | source | |
| Terminal-Bench 2.0 | 49.6% | 8% | Codex CLI | |
| DeepSWE | est. | 6% | predicted 36 | |
| FrontierSWE | est. | 5% | predicted 8 | |
| CursorBench | est. | 4% | predicted 32 | |
| GSO-Bench | 6.9% | 4% | OpenHands | |
| METR Time Horizons | 69.6% | 10% | source | |
| Vending-Bench 2 | est. | 6% | predicted 27 | |
| PostTrainBench | est. | 4% | predicted 21 | |
| APEX-Agents | est. | 8% | predicted 34 | |
| GDPval | 34.8% | 7% | source | |
| Remote Labor Index | 1.7% | 5% | source | |
| DeepResearch Bench | 49.6% | 5% | source | |
| OSWorld 2.0 | est. | 6% | predicted 5 | |
| OSWorld | est. | 4% | predicted 61 | |
| Cybench | est. | 5% | predicted 62 | |
| CL-bench | est. | 5% | predicted 19 |
Other benchmarks
| Benchmark | Category | Score | Source |
|---|---|---|---|
| Aider Polyglot | Coding | 88.0% | source |
| AlgoTune | Coding | 1.67× | source |
| ARC-AGI | Reasoning | 65.7% | source |
| ARC-AGI-2 | Reasoning | 9.9% | source |
| BALROG | Long-horizon | 32.8% | source |
| Chess Puzzles | Reasoning | 36.0% | source |
| CritPt | Reasoning | 12.6% | source |
| DTBench | Knowledge | 90.7% | source |
| Fiction.LiveBench | Long context | 97.2% | source |
| FrontierMath Tiers 1-3 | Math | 55.4% | source |
| FrontierMath Tier 4 | Math | 22.0% | source |
| GeoBench | Multimodal | 81.0% | source |
| GPQA Diamond | Reasoning | 86.2% | source |
| Humanity's Last Exam | Reasoning | 25.3% | source |
| LMCA | Knowledge | 40.0% | source |
| MATH Level 5 | Math | 98.1% | source |
| MirrorCode | Coding | 4.4% | source |
| OTIS Mock AIME | Math | 91.4% | source |
| ProofBench | Math | 18.0% | source |
| SciCode | Coding | 42.9% | source |
| SimpleBench | Reasoning | 56.7% | source |
| SimpleQA Verified | Knowledge | 50.1% | source |
| VPCT | Multimodal | 66.0% | source |
| WebDev Arena | Coding | 1420 | source |