Grok miscounts a long list of teams and keeps defending the answer
ivan_bezdomny · x · 2026-07-23
The thread argues that Grok regressed on a simple but error-prone task: enumerating long lists accurately.
- It reportedly included MLS when it should not have.
- Even after that was removed, it still miscounted teams in Baltimore and Las Vegas.
- The core criticism is that the model kept producing defensive commentary instead of checking its answer carefully.
The reply claims older, “dumber” models handled similar long-list consumer queries better a year ago.
More from Models
- Kimi K3.1 Pro tops MathArena open models, closing frontier gap to about 6 months — teortaxesTex · 2026-10-26
- Ruff creator Charlie Marsh: frontier models find infinite bugs in any software — charliermarsh · 2026-10-26
- Dev testing one skill across Opus 5.5 and Sol 6.1 finds instruction following 'all over the map' — lucasmeijer · 2026-10-26
- supermemory unveils learner-1, claiming continual learning for any model — julianweisser · 2026-10-26
- Anthropic says Opus 5.5 cuts cost per task by ~40% vs Opus 5 — taufiqintech · 2026-10-26
- AI search APIs ranked: top 12 within 7 points, yet prices differ nearly 7x — FinanceYF5 · 2026-10-26