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

Claude Sonnet 4 now supports 1M tokens of context

anthropic.com

Post reply on HN
Print view View original

71–80 of 706 posts

Re: Claude Sonnet 4 now supports 1M tokens of context

#71

Post by film42 »

faangguyindia wrote: post #6

In my testing the gap between claude and gemini pro 2.5 is close. My company is in asia pacific and we can't get access to claude via vertex for some stupid reason. but i tested it via other providers, the gap used to be huge but now not.

Agree but pricing wise, Gemini 2.5 pro wins. Gemini input tokens are half the cost of Claude 4. Output is $5/million cheaper than Claude. But, document processing is significantly cheaper. A 5MB PDF (customer invoice) with Gemini is like 5k tokens vs 56k with Claude.

The only downside with Gemini (and it's a big one) is availability. We get rate limited by their dynamic QoS all the time even if we haven't reached our quota. Our GCP sales rep keeps recommending "provisioned throughput," but it's both expensive, and doesn't fit our workload type. Plus, the VertexAI SDK is kind of a PITA compared to Anthropic.

Re: Claude Sonnet 4 now supports 1M tokens of context

#72

Post by sdesol »

alexchamberlain wrote: post #35
aliljet wrote: post #16

This is definitely one of my CORE problem as I use these tools for "professional software engineering." I really desperately need LLMs to maintain extremely effective context and it's not actually that interesting to see a new model that's marginally better than the next one (for my day-to-day). However. Price is king. Allowing me to flood the context window with my code base is great, but given that the price has su…

I'm not sure how, and maybe some of the coding agents are doing this, but we need to teach the AI to use abstractions, rather than the whole code base for context. We as humans don't hold the whole codebase in our hear, and we shouldn't expect the AI to either.

LLMs (current implementation) are probabilistic so it really needs the actual code to predict the most likely next tokens. Now loading the whole code base can be a problem in itself, since other files may negatively affect the next token.

Re: Claude Sonnet 4 now supports 1M tokens of context

#73

Post by lightbendover »

aliljet wrote: post #27
rootnod3 wrote: post #20

Earlier quoted context omitted.

Flooding the context also means increasing the likelihood of the LLM confusing itself. Mainly because of the longer context. It derails along the way without a reset.

How do you know that?

[dead]

Re: Claude Sonnet 4 now supports 1M tokens of context

#74

Post by mikepurvis »

benterix wrote: post #29
aliljet wrote: post #16

This is definitely one of my CORE problem as I use these tools for "professional software engineering." I really desperately need LLMs to maintain extremely effective context and it's not actually that interesting to see a new model that's marginally better than the next one (for my day-to-day). However. Price is king. Allowing me to flood the context window with my code base is great, but given that the price has su…

> it's not clear if the value actually exists here. Having spent a couple of weeks on Claude Code recently, I arrived to the conclusion that the net value for me from agentic AI is actually negative. I will give it another run in 6-8 months though.

For a bit more nuance, I think I would my overall net is about break even. But I don't take that as "it's not worth it at all, abandon ship" but rather that I need to hone my instinct of what is and is not a good task for AI involvement, and what that involvement should look like.

Throwing together a GHA workflow? Sure, make a ticket, assign it to copilot, check in later to give a little feedback and we're golden. Half a day of labour turned into fifteen minutes.

But there are a lot of tasks that are far too nuanced where trying to take that approach just results in frustration and wasted time. There it's better to rely on editor completion or maybe the chat interface, like "hey I want to do X and Y, what approach makes sense for this?" and treat it like a rubber duck session with a junior colleague.

Re: Claude Sonnet 4 now supports 1M tokens of context

#75

Post by artursapek »

freedomben wrote: post #57
wahnfrieden wrote: post #52

Earlier quoted context omitted.

Did you try with using Opus exclusively?

Do you know if there's a way to force Claude code to do that exclusively? I've found a few env vars online but they don't seem to actually work

You can "force" it by just paying them $200 (which is nothing compared to the value)

Re: Claude Sonnet 4 now supports 1M tokens of context

#76

Post by cambaceres »

benterix wrote: post #29
aliljet wrote: post #16

This is definitely one of my CORE problem as I use these tools for "professional software engineering." I really desperately need LLMs to maintain extremely effective context and it's not actually that interesting to see a new model that's marginally better than the next one (for my day-to-day). However. Price is king. Allowing me to flood the context window with my code base is great, but given that the price has su…

> it's not clear if the value actually exists here. Having spent a couple of weeks on Claude Code recently, I arrived to the conclusion that the net value for me from agentic AI is actually negative. I will give it another run in 6-8 months though.

For me it’s meant a huge increase in productivity, at least 3X.

Since so many claim the opposite, I’m curious to what you do more specifically? I guess different roles/technologies benefit more from agents than others.

I build full stack web applications in node/.net/react, more importantly (I think) is that I work on a small startup and manage 3 applications myself.

Re: Claude Sonnet 4 now supports 1M tokens of context

#78

Post by faangguyindia »

Tostino wrote: post #14
faangguyindia wrote: post #6

In my testing the gap between claude and gemini pro 2.5 is close. My company is in asia pacific and we can't get access to claude via vertex for some stupid reason. but i tested it via other providers, the gap used to be huge but now not.

For me the gap is pretty large (in Gemini Pro 2.5's favor). For reference, the code I am working on is a Spring Boot / (Vaadin) Hilla multi-module project with helm charts for deployment and a separate Python based module for ancillary tasks that were appropriate for it. I've not been able to get any good use out of Sonnet in months now, whereas Gemini Pro 2.5 has (still) been able to grok the project well enough to…

when gemini 2.5 pro gets stuck, i often use deep seek r1 in architect mode and qwen3 in coder mode in aider and it solves all the problem

last month i ran into some wicked dependency bug and only chatgpt could solve it which i am guessing is the case because it has hot data from github?

On the other hand, i really need a tool like aider where i can use various models in "architect" and "coder" mode.

what i've found is better reasoning models tend to be bad at writing actual code, and models like qwen3 coder seems better.

deep seek r1 will not write reliable code but it will reason well and map out the path forward.

i wouldn't be surprised if sonnets success was doing EXACTLY this behind the scenes.

but now i am looking for pure models who do not use this black magic hack behind API.

I want more control at tool end where i can alter the prompts and achieve results i want

this is one reason i do not use claude code etc....

aider is 80% of what i want wish it had more of what i want though.

i just don't know why no one has build a perfect solution to this yet.

Here are things i am missing in aider

1. Automatic model switching, use different models for asking questions about the code, different one for planning a feature, different one for writing actual code.

2. Self determine, if a feature needs a "reasoning" model or coding model will suffice.

3. be able to do more, selectively send context and drop the files we don't need. Intelligently add files to context which will be touched by the feature, not after having done all code planning asking to add files, then doing it all over again with more context available.

Re: Claude Sonnet 4 now supports 1M tokens of context

#79

Post by rootnod3 »

aliljet wrote: post #27
rootnod3 wrote: post #20

Earlier quoted context omitted.

Flooding the context also means increasing the likelihood of the LLM confusing itself. Mainly because of the longer context. It derails along the way without a reset.

How do you know that?

The longer the context and the discussion goes on, the more it can get confused, especially if you have to refine the conversation or code you are building on.

Remember, in its core it's basically a text prediction engine. So the more varying context there is, the more likely it is to make a mess of it.

Short context: conversion leaves the context window and it loses context. Long context: it can mess with the model. So the trick is to strike a balance. But if it's an online models, you have fuck all to control. If it's a local model, you have some say in the parameters.

Post reply on HN