Thread

1. egeres · 1790062027 · focus · HN ↗

   It feels suspicious that MiMo-V2.6 Pro gets 46 in de index while DeepSeek-V4.1 (<a href="https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;models&#x2F;deepseek-v4-1-flash" rel="nofollow">https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;models&#x2F;deepseek-v4-1-flash) gets 39. According to the appendix at the bottom of <a href="https:&#x2F;&#x2F;mimo.xiaomi.com&#x2F;mimo-v2-6" rel="nofollow">https:&#x2F;&#x2F;mimo.xiaomi.com&#x2F;mimo-v2-6 the deepseek model sometimes surpasses mimo and it&#x27;s not so far behind in capabilities. A week ago opus 5 appeared 1 points ahead of fable 5 despite fable being a much smarter model (this has been corrected already)

   1. Shekelphile · 1790091715 · focus · HN ↗

      All of the chinese labs have been overfitting on benchmark data to game the results for a while now - MiMo and Deepseek are not anywhere near frontier and mostly compete with models like Luna - which they are still worse than.

      There isn&#x27;t much compelling reason to use these unless you are just averse to giving money to openai&#x2F;altman. A $20 codex sub gives you ~$150 of luna use per weekly limit, while there isn&#x27;t any good subsidized options for chinese models at all (and the few who were subsidizing, like opencode, rugpulled by reducing monthly limit to $60 to $15 with no notice to users).

      1. nchmy · 1790094116 · focus · HN ↗

         rugpull is a very strong word... I agree that they are not great with their pricing&#x2F;credit communications, but they do state the multipliers quite clearly in various places. And plenty still have $60 or $30.

         1. Shekelphile · 1790094547 · focus · HN ↗

            Changing pricing buried in docs (which they didn&#x27;t actually do when they started switching every model to $15 limit 2 months ago BTW, they only did this once people noticed and called it out) without notifying users is the definition of a rugpull.

Unofficial Hacker News client; not affiliated with Y Combinator.