Kimi Hallu Bench Hard D
AI benchmark · Agentic
What is Kimi Hallu Bench Hard?
An internal Moonshot AI stress test of factual accuracy that measures models' claim-level hallucination rate on deliberately error-prone prompts, run with tools.
General capability mostly
“On Kimi Hallu Bench Hard bench, an internal benchmark designed to stress-test factual accuracy, its claim-level hallucination rate fell from 23.4% for K3 to 8.3%”
kimi.ai/blog/kimi-k3-1-pro ↗
Who built Kimi Hallu Bench Hard
Led by Moonshot AI (Kimi).
Lineage
Benchmarks this one is built on and built from: versions, subsets, extensions and the indexes that combine them.
Built on
Kimi Hallu Bench original
“On Kimi Hallu Bench Hard bench, an internal benchmark designed to stress-test factual accuracy, its claim-level hallucination rate fell from 23.4% for K3 to 8.3%”
kimi.ai/blog/kimi-k3-1-pro ↗
Tier D
Long tail: little or no usage data yet.
How much stands behind a tier: usage, who backs it, a confirmed line from its own site or paper, and a clear read on X. High with three of the four, medium with two, low otherwise.
- no usage data
- in curated RL directories (1 pt)
On X · unclear, low buzz, from 1 post about Kimi Hallu Bench Hard in the last month
Mentioned only in passing as an internal Moonshot benchmark in Kimi K3.1 Pro launch coverage, with no independent discussion or adoption.
How we got to tier D
- Adoption: no usage data. Tier D.
- On X: unclear, low buzz. “Mentioned only in passing as an internal Moonshot benchmark in Kimi K3.1 Pro launch coverage, with no independent discussion or adoption.” No move.
- In curated RL directories: 1 point, a floor of none.
- Confirmed: a line from its own site or paper places it.
A clear view on X moves a tier one step up or down; most entries don't move. How tiers work.
Where Kimi Hallu Bench Hard sits
Its domain on the map, live. Hover a logo for its name; click to open it. Open it on the full map
Next: browse all benchmarks. Entry from the RL Research daily scrape of public sources, 2026-10-29 snapshot.