GPT Sol 6.1 tops long-horizon benchmark at 50.45%, best open model at 11.86%
teortaxesTex · x · 2026-11-07
According to a cited benchmark update, GPT Sol 6.1 now uses the 12-hour time frame more consistently and tops the benchmark at 50.45%, at a higher cost than Opus 5.5. Open models lag considerably: the best open model, Kimi K3.1 Pro, scores only 11.86%. The poster notes this is another fresh task the frontier is already strong at.
More from Models
- Uncensored DeepSeek-distilled Qwen3.5-9B merge hits HF trending — xtools-at · 2026-11-07
- Anthropic paper: 3M tokens of moral dilemmas cut model malicious behavior to 0%, 28x efficiency — garrytan · 2026-11-07
- Training a Local Qwen 9B with a Frontier Model Boosted Test Scores 3x — SouthernRow2710 · 2026-11-07
- Could CVE rates set a security bar for code-generating models? — _arohan_ · 2026-11-07
- PFN open-sources PLaMo 3 Translate: 0.8B model beats rivals 30x its size — pmttyji · 2026-11-07
- Charted: how open-model parameter counts have evolved by company over time — Independent_Fall9160 · 2026-11-07