Skip to content
View original post on X: vLLMOfficial
Published
26/100

Kimi K3.1 Pro scores 61.2% on Kimi Code Bench 3.0 long-horizon tasks

AIMoonshot AI's K3.1 Pro scores 61.2% on Kimi Code Bench 3.0 Long-horizon, up from K3's 55.95%, according to a post from vLLM. The internal benchmark covers multi-step execution, tool use and end-to-end engineering, and both coding results use the Claude Code harness at maximum reasoning effort.
vLLMVerified on XOfficial
@vllm_project
View on X

Part of a thread · earlier post

Long tasks are the point.

K3.1 Pro scores 61.2% on Kimi Code Bench 3.0 Long-horizon, up from K3’s 55.95%. This internal benchmark covers multi-step execution, tool use and end-to-end engineering.

Both coding results use the Claude Code harness at maximum reasoning effort.

Source details

vLLM · x.comPublished · added here

Original title: Long tasks are the point.

Topics

Kimi / Moonshot AIAI codingReasoningChina AIEvals & benchmarks