Marco Polo Benchmark Shows Chinese Model Retention Falling, Sparking Questions
teortaxesTex · x · 2026-09-23
- teortaxesTex flags an odd trend: Chinese models' retention on the Marco Polo benchmark appears to be falling.
- The author posts data screenshots pointing out several inconsistencies in the results and asks why — prompting debate over how Chinese models' long-context performance is actually being evaluated.
Related event: Marco Polo Data Shows Chinese AI Research Output Up, US Retention Down(3 posts)→
More from Models
- YC-backed Monte launches post-training platform to build proprietary models from your agents' real work — ycombinator · 2026-11-12
- Leak Claims OpenAI Will Launch Entire GPT-6.2 Family Tomorrow to Counter Opus 5.5 and Fable 5.2 — heypearlai · 2026-11-12
- ARC-AGI-3 leaderboard: models score 90%+ but at $10K+ per task — haider1 · 2026-11-12
- LMArena Launches Agent Arena for Head-to-Head Agent Evaluation — arena · 2026-11-12
- LLMs lack hyperlocal municipal data, explains disputed Texas power outage debate — csuwildcat · 2026-11-12
- Claude API's per-message effort lets you switch reasoning levels without breaking prompt cache — lyraxana · 2026-11-12