Reasoning & Test-Time Compute · 2025

Kimi k1.5: Scaling Reinforcement Learning with LLMs

Kimi Team

Presents a full RL recipe for a frontier reasoning model built on long-context scaling and simplified policy optimization without value functions, MCTS, or process reward models.

Editorial record

Plain-language summary

Presents a full RL recipe for a frontier reasoning model built on long-context scaling and simplified policy optimization without value functions, MCTS, or process reward models.

Knowledge graph

Relationships

Descendants

Source record

Provenance

Record ID
P-637
Record created
2026-07-13
Last reviewed
2026-07-14
Record version
2

Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.