Thread

1. GodelNumbering · 1790095446 · focus · HN ↗

   Finally that price drop

      Prices per 1M tokens     Claude Opus 5.5    Claude Opus 5
      Cache reads              $0.20              $0.50
      Input tokens             $4                 $5
      Output tokens            $20                $25
      Cache writes             $5                 $6.25
   
   

   Opus 5 is the model with highest spend on openrouter (<a href="https:&#x2F;&#x2F;openrouter.ai&#x2F;rankings#task-spend" rel="nofollow">https:&#x2F;&#x2F;openrouter.ai&#x2F;rankings#task-spend) and it seems plausible that Opus 5 is&#x2F;was the highest spend model in the world, and certainly Anthropic&#x27;s biggest moneymaker.

   If you are forced to reduce price despite raising capabilities, that certainly tells something about the market, and potentially about Anthropic future profitability too, since this model is their biggest topline contributor

   1. AJ007 · 1790096652 · focus · HN ↗

      It is only a price drop if price * tokens used is less

      1. mcintyre1994 · 1790097162 · focus · HN ↗

         They&#x27;re claiming a drop in token use too, and that it nets to 40% cheaper.

         1. drbscl · 1790097387 · focus · HN ↗

            Unfortunately, they&#x27;re full of it <a href="https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;models&#x2F;claude-opus-5-5#token-use" rel="nofollow">https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;models&#x2F;claude-opus-5-5#token-u...

            It does work out to be a similar cost per task though

            1. jsnell · 1790097862 · focus · HN ↗

               You should probably look at the cost&#x2F;score graph by effort level instead:

               <a href="https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;models&#x2F;claude-opus-5-5#intelligence-comparisons" rel="nofollow">https:&#x2F;&#x2F;artificialanalysis.ai&#x2F;models&#x2F;claude-opus-5-5#intelli...

               It is most of the pareto frontier.

               1. drbscl · 1790098195 · focus · HN ↗

                  Not disputing the increase in quality, just stating that non-cherry-picked benchmarks show it is more verbose at Max effort

                  1. 93po · 1790101294 · focus · HN ↗

                     Is verboseness the only measure of token efficiency towards overall task completion?
                  2. persedes · 1790108363 · focus · HN ↗

                     so don&#x27;t use it at max? The benchmarks suggest that high&#x2F;xhigh are more than sufficient to be ahead and a whole magnitude below max with regards to token usage. I&#x27;d treat that as an outlier and not how verbose the model is in general (QED I know)

                     1. drbscl · 1790148146 · focus · HN ↗

                        You’re missing my point. I’m saying anthropic are exaggerating their results.

                        1. persedes · 1790179318 · focus · HN ↗

                           how are they exaggerating the results? Comparing the cost from that chart for 5 and 5.5 for medium-max effort paints a pretty clear picture:

                                    mean  median
                            model
                            5      4.135   4.245
                            5.5    3.150   2.640
                           
                           

                           Again seeing how max is a clear outlier, the median cost saving is ~38%, not that far off from the proclaimed 40%.

Unofficial Hacker News client; not affiliated with Y Combinator.