Sharper GPT-5 generation for coding, product work, and tool-assisted tasks
Benchmarks
| Benchmark | Category | Score | Variant | Provenance | Source | Observed |
|---|---|---|---|---|---|---|
| AA Agentic Index | agentic | 21.0 | — | mirror | openrouter | Jul 23, 2026 |
| 21.0 | — | independent | aa | Jul 24, 2026 | ||
| AA Coding Index | coding | 49.4 | — | mirror | openrouter | Jul 23, 2026 |
| 49.4 | — | independent | aa | Jul 24, 2026 | ||
| AA Intelligence Index | intelligence | 36.9 | — | mirror | openrouter | Jul 23, 2026 |
| 36.9 | — | independent | aa | Jul 24, 2026 | ||
| Design Arena — 3d | preference | 1121 | — | mirror | openrouter | Jul 24, 2026 |
| Design Arena — asciiart | preference | 1159 | — | mirror | openrouter | Jul 23, 2026 |
| Design Arena — codecategories | preference | 1204 | — | mirror | openrouter | Jul 23, 2026 |
| Design Arena — dataviz | preference | 1238 | — | mirror | openrouter | Jul 24, 2026 |
| Design Arena — gamedev | preference | 1237 | — | mirror | openrouter | Jul 23, 2026 |
| Design Arena — mobileapps | preference | 1123 | — | mirror | openrouter | Jul 23, 2026 |
| Design Arena — svg | preference | 1198 | — | mirror | openrouter | Jul 23, 2026 |
| Design Arena — uicomponent | preference | 1207 | — | mirror | openrouter | Jul 24, 2026 |
| Design Arena — webapps | preference | 1075 | — | mirror | openrouter | Jul 23, 2026 |
| Design Arena — website | preference | 1215 | — | mirror | openrouter | Jul 23, 2026 |
| LMArena Text | preference | 1485 | — | crowd | lmarena | Jul 21, 2026 |
| 1482 | high-effort | crowd | lmarena | Jul 21, 2026 | ||
| LMArena Vision | preference | 1248 | high-effort | crowd | lmarena | Jul 21, 2026 |
| 1232 | — | crowd | lmarena | Jul 21, 2026 | ||
| LMArena WebDev | preference | 1339 | — | crowd | lmarena | Jul 21, 2026 |
| Median output speed | speed | 88 tok/s | — | independent | aa | Jul 24, 2026 |
| Median time to first token | speed | 47.62s | — | independent | aa | Jul 24, 2026 |
Includes index data from Artificial Analysis.
Providers
Live operational stats per hosting endpoint, via OpenRouter.
| Provider | Quant | Context | $/1M in · out | Uptime 24h | Latency | Throughput |
|---|---|---|---|---|---|---|
| Azure | — | 400K | $1.25 · $10 | 99.94% | — | — |
| Azure | — | 400K | $1.38 · $11 | — | — | — |
| OpenAI | — | 400K | $1.25 · $10 | 94.33% | — | — |
| OpenAI | — | 400K | $0.625 · $5 | 94.33% | — | — |
| OpenAI | — | 400K | $2.5 · $20 | 94.33% | — | — |