As of 2026-10-08
Other desks.
Third-party roundups.
Fello, Vals, Santage, Local AI Master, Developers Digest. Rank follows WebDev Arena Elo when present, else SWE Verified.
Roundups #1 is Opus 5
Other desks
Third-party roundups. Coverage is thinner than AA, Arena, or BenchLM.
| 1 | Opus 5 Anthropic · closed Anthropic’s flagship frontier model | 1702.9Roundup | — | 1858Lab |
| 2 | Kimi K3 Moonshot AI · open Frontier open-weight | 1675.5Roundup | 93.4%Roundup | — |
| — | Fable 5 Anthropic · closed New Claude 5 frontier tier | — | — | 1746Lab |
| 3 | GPT-5.6 Sol OpenAI · closed Top-tier flagship reasoning | — | 96.2%Lab | — |
| — | Spark 1.2 Meta · closed Agent-native Meta model | — | — | 1631Roundup |
| 4 | Grok 4.5 SpaceXAI · closed Frontier coding/agents | — | 86.6%Roundup | — |
6 models on this board · # is this page’s rank