... here. Frontier models sometimes scored lower because they were too smart, adding guideline level recommendations instead of extracting what the note actually said. The finetuned model better matched the expected extraction style, correctly distinguishing "yes" for chronic conditions vs "no" for routine procedures in follow-up flags. Legal Clause Rewriting (Qwen3-32B) : Tied with one of the top frontier models, within 0.25 of the best. The finetuned model learned to explicitly restate each legal qualifier in simple terms rather than ... ... ways to finetune nowadays. Here's the comparison. OpenAI finetuning is polished and easy but you can't download your model. It lives on their servers, you pay per token forever, and if they deprecate it, it's gone. Unsloth

... subscriptions no longer cover usage on third-party tools like @OpenClaw . If you were running your agent with @Claude by Anthropic , curious what option you'd pick: would you pay the extra bundle? pick another provider? or use an alternative like @KiloClaw or @Zo Computer ? Comment from Sai Tharun Kakirala(@sai_tharun_kakirala): This was a rough change for a lot of builders who had workflows running on top of the Claude subscription model. The "free tier via subscription" loophole ... ... teams will rediscover that the model choice and the execution environment should be evaluated separately. Sometimes the cheapest model is not the cheapest workflow once you include retries, weak guardrails, or bad ergonomics. Would be interesting to hear which alternatives