‹ BackHN Continuity

Thread

GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price

1066 points · 953 comments · crorella

  1. phpnode · · focus · HN ↗
    What's driving the increase in release cadence here? We seem to get new models every week or so now, is this RSI?
    1. sharpshadow · · focus · HN ↗
      Response to DeepSeek’s technical paper and competition.
      1. wg0 · · focus · HN ↗
        What's that in summary?
        1. Wheen · · focus · HN ↗
          Not the person you're replying to, but judging by the emphasis on the cost of cached input tokens in the OP article, I'd guess it has to do with DeepSeek v4.1's KV cache efficiency. It uses <1000 bytes per token, so they're able to get 1M token context in under a GB.

          Edit: <a href="https:&#x2F;&#x2F;huggingface.co&#x2F;deepseek-ai&#x2F;DeepSeek-V4.1-Flash&#x2F;blob&#x2F;main&#x2F;DeepSeek_V41_Tech_Report.pdf" rel="nofollow">https:&#x2F;&#x2F;huggingface.co&#x2F;deepseek-ai&#x2F;DeepSeek-V4.1-Flash&#x2F;blob&#x2F;...

          1. ChromeUltron · · focus · HN ↗
            just goes to show that OpenAI in fact did not innovate on a single thing for the better part of a year (one could argue two) and instead keeps immitating what it sees doing others successfully with the tech, all in a very transparent attempt to get people lubed up for their IPO.
      2. LPisGood · · focus · HN ↗
        Which paper are you referring to?
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.