PI-OSBenchmarks
← PI-OS Benchmarks / Deal Screener

Model Profile: Kimi K3

Candidate Card

Shareable BS-0 native evidence summary for Kimi K3 — outcome counts, median TTFT, generation throughput, provider tokens and the DM-07 Silence Trap result, as captured in the frozen native run.

BS-0 native candidate card — Kimi K3 — Evidence-Disciplined Judge; 6 aligned, 0 partial, 1 misaligned. DM-07 Silence Trap failed.
OBSERVED — BS-0 native (un-governed) evidence snapshot. Governed PI-OS qualification: NOT YET RUN. Candidate glyphs are PI-SF evidence marks generated for this study — they are not official vendor logos and imply no vendor endorsement.Download card (PNG)

At a Glance

ProviderMoonshot AI
Model SizeNot captured
Context WindowNot captured
SettingNative (Un-governed)
Inference SettingsT=0, Top P=1, Top K=40
Runs Completed7 / 7
Canonical Observations7 / 7
TTFT (median)1,054 ms observed
Generation Throughput (median)57.10 tok/s observed
Provider Tokens (run total)19.1K observed
Native Outcomes (A / P / M)6 / 0 / 1

Governance Fingerprint

RepairDetectSuppress

Lower area = lower governance burden

low-repair / low-detect / low-suppress (derived from case outcomes)

Case Results (Native)

Strengths

  • Most aligned outcomes observed (6 of 7)
  • No partial outcomes recorded

Watchouts

  • Slowest observed generation throughput (57.10 tok/s)
  • Fails closed-world exclusion (DM-07)
  • Dimension grades derived from case outcomes, not separately graded

49 / 49 observations · dataset frozen 25 August 2026