Kimi K3.1 Pro Official Evaluation ReportProvider report

Name: Kimi K3.1 Pro Official Evaluation Report
Published: 2026-10-24

2026-10-24Kimi K3.1 Pro only: 47 release-table rows become 48 metrics after splitting ZeroBench avg@5 and pass@5; two finance-chart results bring the total to 50. Versions, tools, scoring rules and internal benchmarks remain explicit. Vendor-reported results at max effort unless stated otherwise. Business Operation uses the main table value 37.8; the finance chart reports 37.0. Highspeed shares the model's capabilities and is not scored separately.

Back to data sources

Kimi K3.1 Pro Official Evaluation ReportProvider report

View source

2026-10-24Kimi K3.1 Pro only: 47 release-table rows become 48 metrics after splitting ZeroBench avg@5 and pass@5; two finance-chart results bring the total to 50. Versions, tools, scoring rules and internal benchmarks remain explicit. Vendor-reported results at max effort unless stated otherwise. Business Operation uses the main table value 37.8; the finance chart reports 37.0. Highspeed shares the model's capabilities and is not scored separately.