manton
manton

Interesting note from Simon Willison that Claude’s new tokenizer translates to higher costs:

Opus 4.7 uses the same pricing is Opus 4.6 - $5 per million input tokens and $25 per million output tokens - but this token inflation means we can expect it to be around 40% more expensive

Anthropic pricing is already much more expensive than OpenAI. With how good the new Codex app for Mac is, I think we’ll see more developers move to Codex. But people are also comfortable with their workflows, so there’s not going to be a quick shift, and models and features change nearly every month.

|
Embed
Progress spinner
numericcitizen
numericcitizen

@manton but a one year subscription with Claude on January 1st. 😳 I’ll keep using Claude but might switch near the end of the year, especially if OpenAI really gets better and if tokens consumption issue isn’t fixed by then.

|
Embed
Progress spinner
manton
manton

@numericcitizen It’ll be interesting to watch. I think Anthropic is constrained by compute, but that’ll improve with time as they scale servers. Also maybe ironically if some people switch away, that also frees up resources, making it faster / better for others!

|
Embed
Progress spinner
In reply to
ryan@goodlett.social
ryan@goodlett.social

@manton I can tell you at my employer there is a 7.5x multiplier on 4.7 Opus vs a 3x on 4.6 Opus via MS Copilot.

|
Embed
Progress spinner