Kimi K3.1 Pro costs 52x more with parity success on Runescape Bench, new checkpoint suspected
teortaxesTex · x · 2026-11-06
Researcher teortaxesTex criticizes Kimi K3.1 Pro's Runescape Bench results: 52x more expensive than DS V4.1 Flash 1017 with more tokens but parity in success rate — 'bad bets.' When others blamed inference providers, the author counters that this is an entirely new checkpoint, not the original K3.1 Pro: Moonshot did DSA adaptation within the last two weeks and healed it on new tasks. The attached chart breaks down per-skill XP/min across models, with Kimi K3.1 Pro mid-pack overall.
More from Models
- Scale AI ships 3,300+ RL environments; some tasks take humans 3 weeks and 500+ tool calls — alexandr_wang · 2026-11-06
- Mercury 3 hits 2 trillion tokens in under two weeks as daily usage climbs linearly — StanfordAILab · 2026-11-06
- Same local model gives different answers via LM Link on iPhone vs direct on Mac — simplext · 2026-11-06
- "Anthropic valuation dropped from $2T to $20B — AI told me" — michalmalewicz · 2026-11-06
- One-letter URL encoding bypass exposed Hugging Face chat admin API, bulk export of private chats — cyb3rops · 2026-11-06
- Tencent Hy4 lands on OpenDesign with frontend and complex UI upgrades — TencentHunyuan · 2026-11-06