Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Measuring Claude 4.7's tokenizer costs

claudecodecamp.com

Post reply on HN
Print view View original

181–190 of 540 posts

Re: Measuring Claude 4.7's tokenizer costs

#181

Post by throwaway041207 »

margorczynski wrote: post #138

It doesn't look good for Anthropic, especially considering they are burning billions in investor money. Looks like they lost the mandate of heaven, if Open AI plays it right it might be their end. Add to that the open source models from China.

I work at a company that has gone all in on Anthropic, and we're just shoveling money at them. I suspect there are a more enterprises than we realize that are doing this.

When I read these comments on Hacker News, I see a lot of people miffed about their personal subscription limits. I think this is a viewpoint that is very consumer focused, and probably within Anthropic they're seeing buckets of money being dumped on them from enterprises. They probably don't really care as much about the individual subscription user, especially power users.

Re: Measuring Claude 4.7's tokenizer costs

#182

Post by __mharrison__ »

beej71 wrote: post #65

News like this always makes me wonder about running my own model, something I've never done. A couple thousand bucks can get you some decent hardware, it looks like, but is it good for coding? What is your all's experience? And if it's not good enough for coding, what kind of money, if any, would make it good enough?

My anecdotal experience with a recent project (Python library implemented and released to pypi).

I took the plan that I used from Codex and handed it to opencode with Qwen 3.5 running locally.

It created a library very similar to Codex but took 2x longer.

I haven't tried Qwen 3.6 but I hear it's another improvement. I'm confident with my AI skills that if/when cheap/subsidized models go away, I'll be fine running locally.

Re: Measuring Claude 4.7's tokenizer costs

#183

Post by mgraczyk »

_pdp_ wrote: post #29

IMHO there is a point where incremental model quality will hit diminishing returns. It is like comparing an 8K display to a 16K display because at normal viewing distance, the difference is imperceptible, but 16K comes at significant premium. The same applies to intelligence. Sure, some users might register a meaningful bump, but if 99% can't tell the difference in their day-to-day work, does it matter? A 20-30% cost…

This will probably happen but I wouldn't plan on it happening soon

Re: Measuring Claude 4.7's tokenizer costs

#184

Post by GaryBluto »

Someone1234 wrote: post #113
speedgoose wrote: post #103

The "multiplier" on Github Copilot went from 3 to 7.5. Nice to see that it is actually only 20-30% and Microsoft wanting to lose money slightly slower. https://docs.github.com/fr/copilot/reference/ai-models/suppo...

Yep, and I just made a recommendation that was essentially "never enable Opus 4.7" to my org as a direct result. We have Opus 4.6 (3x) and Opus 4.5 (3x) enabled currently. They are worth it for planning . At 7.5x for 4.7, heck no. It isn't even clear it is an upgrade over Opus 4.6.

Microsoft are going to be removing Opus 4.5 and 4.6 from Copilot soon so I'd enjoy the lower cost while it lasts.

Re: Measuring Claude 4.7's tokenizer costs

#185

Post by Bridged7756 »

_fat_santa wrote: post #62

A question I've been asking alot lately (really since the release of GPT-5.3) is "do I really need the more powerful model"? I think a big issue with the industry right now is it's constantly chasing higher performing models and that comes at the cost of everything else. What I would love to see in the next few years is all these frontier AI labs go from just trying to create the most powerful model at any cost to ac…

Efficiency doesn't make as much money. It's in big LLM's best interest to keep inference computationally expensive.

I personally think the whole "the newest model is crazy! You've gotta use X (insert most expensive model)" Is just FOMO and marketing-prone people just parroting whatever they've seen in the news or online.

Re: Measuring Claude 4.7's tokenizer costs

#186

Post by cbg0 »

solenoid0937 wrote: post #159
cbg0 wrote: post #15

Earlier quoted context omitted.

That performance monitor is super easy to game if you cache responses to all the SWE bench questions.

You dramatically overestimate how much time engineers at hypergrowth startups have on their hands

Caching some data is time consuming? They can just ask Claude to do it.

Re: Measuring Claude 4.7's tokenizer costs

#187

Post by cced »

bisonbear wrote: post #161
namnnumbr wrote: post #56

The title is a misdirection. The token counts may be higher, but the cost-per-task may not be for a given intelligence level. Need to wait to see Artificial Analysis' Intelligence Index run for this, or some other independent per-task cost analysis. The final calculation assumes that Opus 4.7 uses the exact same trajectory + reasoning output as Opus 4.6. I have not verified, but I assume it not to be the case, given…

yep, ran a controlled experiment on 28 tasks comparing old opus 4.6 vs new opus 4.6 vs 4.7, and found that 4.7 is comparable in cost to old 4.6, and ~20% more expensive then new 4.6 (because new 4.6 is thinking less) https://www.stet.sh/blog/opus-4-7-zod

So they nerfed 4.6 to make way for 4.7?

Progress. /s

Re: Measuring Claude 4.7's tokenizer costs

#188

Post by bisonbear »

cced wrote: post #187
bisonbear wrote: post #161

Earlier quoted context omitted.

yep, ran a controlled experiment on 28 tasks comparing old opus 4.6 vs new opus 4.6 vs 4.7, and found that 4.7 is comparable in cost to old 4.6, and ~20% more expensive then new 4.6 (because new 4.6 is thinking less) https://www.stet.sh/blog/opus-4-7-zod

So they nerfed 4.6 to make way for 4.7? Progress. /s

> they nerfed 4.6 to make way for 4.7?

> Progress. /s

pretty much, lmao. my theory is 4.6 started thinking less to save compute for 4.7 release. but who knows what's going on at anthropic

Re: Measuring Claude 4.7's tokenizer costs

#189

Post by AndyNemmity »

aray07 wrote: post #88
aliljet wrote: post #79

This is the reality I'm seeing too. Does this mean that the subscriptions (5x, 10x, 20x) are essentially reduced in token-count by 20-30%?

yeah thats the part that is unclear to me as well - if our usage capacity is now going to run out faster.

The same thing I've been doing all the time, now has used up 1/3rd of my week in one day on max20.

So yes, for the same tasks, usage runs out faster (currently)

Re: Measuring Claude 4.7's tokenizer costs

#190

Post by manmal »

UncleOxidant wrote: post #95
ZeroCool2u wrote: post #71

Earlier quoted context omitted.

Whenever we get the locally runnable 4k models things are going to get really awkward for the big 3 labs. Well at least Google will still have their ad revenue I guess.

Given how little claude usage they've been giving us on the "pro" plan lately, I've started doing more with the various open Qwen3.* models. Both Qwen3-coder-next and Qwen3.5-27b have been giving me good results and their 3.6 models are starting to be released. I think Anthropic may be shooting themselves in the foot here as more people start moving to local models due to costs and/or availability. Are the Qwen model…

Why don’t you do the planning yourself? It’s very likely to be a better plan.
Post reply on HN