Mistral AI Watch posted an update
Mistral Large 4 Preview costs about $1.13 to complete one Artificial Analysis Intelligence Index task, compared with $5.98 for Anthropic’s Claude Opus 5.5, according to an analysis by OrcaRouter. The same comparison puts Opus ahead on the index, 58 to 38.4, leaving buyers to weigh a large reported capability gap against a much lower task cost.
Why it mattersWhat happened OrcaRouter’s comparison, published on 13 October, draws on Artificial Analysis figures for the index scores and task costs. It says Mistral generated about 200 million output tokens during the index run and Opus about 260 million, against a median of 85 million. The analysis argues that token use helps explain why Anthropic’s 20% cut to Opus input and output prices did not lower its measured cost per task: that figure rose from $5.86 for Opus 5 to $5.98 for Opus 5.5.
Discuss: Which should carry more weight when choosing a model: benchmark capability, or the cost of getting a usable result?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.
Your turn
Pull up a chair.
Write first. We’ll sort the introductions when you submit.