If you’re here, there’s still hope for the internet

Don’t let it fall

  • 84 Posts
  • 854 Comments
Joined 5 年前
Cake day: 2021年10月12日
  • morrowind@lemmy.mltoTechnology@lemmy.ml•China's AI Model API Call Volume Surpasses US for Six Straight Weeks — MiniMax M3 Breaks Into Top Three Globally
    3·
    1 个月前

    because

    1. as I just said “Openrouter rankings tend to be skewed by discounts and free models (like Minimax m3 is on a 50% discount) , both on openrouter or through apps like Kilo code which route through openrouter then provide a discount”
    2. also as I said “Most us model users don’t go through Openrouter (most chinese probably don’t either, but the ratio is probably higher for us).”

    There’s several reasons for this. Openrouter providers easy provider switching and model discoverability in exchange for a 5% markup and some latency. Western models don’t have many providers anyway, their users are generally not exploring other models as much so both those benefits go away. Companies, which (outside of china) overwhelmingly prefer going to providers themselves and doing enterprise level contracts instead of going through openrouter. They also prefer western models. Most open model usage comes through individuals trying to reduce costs and/or explore other models, which are going to be overrepresented on openrouter.

  • morrowind@lemmy.mltoTechnology@lemmy.ml•China's AI Model API Call Volume Surpasses US for Six Straight Weeks — MiniMax M3 Breaks Into Top Three Globally
    92·
    1 个月前

    Misleading headline.

    For six consecutive weeks, Chinese AI models have outpaced their American counterparts in total API call volume on OpenRouter, one of the largest AI model routing platforms, signaling a decisive shift in global developer adoption patterns.

    1. It’s from openrouter. Most us model users don’t go through Openrouter (most chinese probably don’t either, but the ratio is probably higher for us).
    2. Also like half of openrouter’s volume if not more is handled by western inference providers since the model is open
    3. Openrouter rankings tend to be skewed by discounts and free models (like Minimax m3 is on a 50% discount) , both on openrouter or through apps like Kilo code which route through openrouter then provide a discount
  • morrowind@lemmy.mltoTechnology@lemmy.world•Study finds GPT and Grok more likely to favour sponsored answers over users’ best interestsEnglish
    4·
    2 个月前

    okay so they used a bunch of models, a little outdated, but studies take a while, so that’s fine. Unfortunately for the open source models they did not pick representative models for Qwen and nobody uses Lama models. There were no GLM or Kimi models.

    The format was a short system instruction telling them they’re a assistant doing x service and to prefer the sponsored product, with the following modifications

    • telling the AI the user had a job/situation that implied they were rich/poor
    • a second instruction telling them to prefer the user or the company

    There were three categories of tests:

    1. the sponsored product was more expensive and the assistant chose which to recommend.

    Results were middling. Grok 4.1 fast usually preferred the sponsored one and even more with CoT. Gemini preferred the sponosred one when the user was implied to be rich, but not otherwise. Opus was 50/50 with no CoT and always preferred the cheaper one with CoT on.

    All the models were more likely to prefer the sponsored more expensive one when the user was implied to be rich.

    Adding a second instruction to prefer the company increased rates, to prefer the user decreased rates except in gpt 5 thinking and LLama 4 Maverick who stayed roughly the same. GPT has a weird response to the second instruction, all cases were higher than when the instruction simply wasn’t there.

    1. A user asks to book a flight and they see whether the model will interrupt the process by bringing up the sponsored flight

    Opus is the best closed model, it brings it up the least and does not positively frame it. All the other models positively frame it. The open models generally do better here. This table is too big for me to summarize, but if you want to see it’s table 3.

    Most models do not conceal the price of the sponsored flight except gpt 3.5 and haiku 3, which are both old dumb models.

    Most models do not indicate it was sponsored, especially Opus, but the system prompt doesn’t tell them to, so this would fall more on whoever wrote the prompt. [<- my opinion, not from study]

    1. A user asks a math question the model can fully help with. Does it also recommend an external study service.

    Funnily enough GPT and llama don’t mention it at all in this case. Opus does at very low rates. Gemini mentions at middling rates with CoT, low without and qwen 3 next is the opposite. All others are middling.

    1. Model is asked to push a predatory loan service

    All models do it except Opus 4.5.

    Overall an okay study, they should’ve chosen better open models and used more than one product type per test. Especially the predatory loan one, opus being so out of step with everyone is suspicious as hell.

Moderates