Kimi K3.1 Pro Scores 0.78x Compute Multiplier in TasteVal Benchmark

pzeroresearch's TasteVal benchmark measures Kimi K3.1 Pro at a 0.78x experimental compute multiplier versus human experts, about 4.6 months behind the frontier; it performs strongly under low compute budgets but gains drop with more compute.

2026-11-05 ~ 2026-11-05 · 2 related posts

• Kimi K3.1 Pro hits 0.78x human compute multiplier, lagging frontier by ~4.6 months — ChowdhuryNeil · 2026-11-05
• Kimi K3.1 Pro gains less from test-time compute than Opus 5.0, data shows — teortaxesTex · 2026-11-05