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

GPT-5.6

openai.com

Post reply on HN
Print view View original

751–760 of 1001 posts

Re: GPT-5.6

#751

Post by noisy_boy »

davidhs wrote: post #339
Syntaf wrote: post #6

Ok long time Claude Code user here; lately I've started to realize there's other great models out there I should be trying, but I'm hesitant to leave Claude Code behind for something new. What's the consensus today on codex vs claude code, does it really matter anymore?

I recommend trying Codex too. In fact, I recommend running them side-by-side if you have the budget, e.g. have both independently plan the same feature or implement in a different worktree, or have them critique each other's work. I personally find GPT-5.5 to be a better programmer than Opus 4.8, it is extremely thorough, but I don't like the code it generates ("austere"), and find Opus 4.8 to write more "human frien…

I did the side by side between Claude code (effort medium) and Cursor (auto). Asked Claude to prepare a plan and asked Cursor to review it and it found tons of gaps in the plan. Cost-wise, it came out better too. I have been using Cursor daily (along with Claude) and the former has been 20-30% cheaper despite me spending more time on it.

Re: GPT-5.6

#752

Post by Sabinus »

doodlesarefun wrote: post #748
user43928 wrote: post #634

Earlier quoted context omitted.

I let it work on two features while I was using mostly GPT 5.6 and it has already consumed 10% of the weekly Fable limit on Max x20. GPT 5.6 on the Pro x5 plan is down to... 100%. It looks like they just reset the usage limits again. And I still have two resets on the bench. Anthropic is going to have to up their game to compete.

What do you mean 'resets on the bench?

When OpenAI do quota resets (not your scheduled weekly ones, the promotional or apology resets) they don't just reset whatever quota you were on. They give you a token to reset your weekly quota when you like.

Re: GPT-5.6

#753

Post by sunnybeetroot »

fblp wrote: post #562
eig wrote: post #38

Funny to see that they did not include Fable 5 in their GeneBench and LifeSciBench comparisons because "it does not answer advanced biology questions and refuses the majority of questions in this eval". Winner by default!

oh that's sad, are the biolgy limitations for "safety"?

Yes

Re: GPT-5.6

#754

Post by y1n0 »

echelon wrote: post #702
navigate8310 wrote: post #515

Earlier quoted context omitted.

There goes my plan to buy a PC for the next decade

The whole of knowledge work is being automated. We've barely begun to see the GPU build out. This is just the start. I'd imagine they're going to 10x this, maybe 100x this.

yeah, spacex's planned million satellite datacenter-in-space constellation doesn't seem as absurd.

Re: GPT-5.6

#755

Post by y1n0 »

Anybody have an idea of what the flops per token generated is on a SOTA model like current GPT/OPUS? Is it basically the parameter count? So something like GLM-5.2 is, at a minimum, ~744 GFLOPs per token generated?

Am I way off base? Seems astronomical.

Re: GPT-5.6

#756

Post by sidgarimella »

I've found Sol's propensity for delegating to subagents can make it... disastrously expensive, especially with each subagent having some implicit floor on further reasoning/context gathering before action.

The base model is certainly cheaper and more token efficient etc, but on large tasks cost in some way is now n^2

Re: GPT-5.6

#757

Post by winwang »

sidgarimella wrote: post #756

I've found Sol's propensity for delegating to subagents can make it... disastrously expensive, especially with each subagent having some implicit floor on further reasoning/context gathering before action. The base model is certainly cheaper and more token efficient etc, but on large tasks cost in some way is now n^2

How bad is it for you? Are you on ultra or xhigh/max? I typically ask it (5.5, now 5.6-sol) to use subagents for specific things anyway. On the Pro 20x plan, I'm seeing like ~1% usage per 20-30 min per session (on max effort), which is in line with 5.5. Currently trying out ultra on a personal project, feels like ~3x more expensive per unit time. (No idea on quality yet, for obvious reasons.)

Re: GPT-5.6

#758

Post by alasano »

sidgarimella wrote: post #756

I've found Sol's propensity for delegating to subagents can make it... disastrously expensive, especially with each subagent having some implicit floor on further reasoning/context gathering before action. The base model is certainly cheaper and more token efficient etc, but on large tasks cost in some way is now n^2

My global CLAUDE.md/AGENTS.md etc have strict rules to never use subagents without approval or when specifically requested except for whitelisted scenarios.

I think that's the way to go most of the time

Re: GPT-5.6

#759

Post by RossBencina »

ravenstine wrote: post #94
minimaxir wrote: post #26

The developer's guide ( https://developers.openai.com/api/docs/guides/latest-model ) has some interesting semantic tips for using the model: > Intent understanding: GPT-5.6 can better infer the user’s underlying goal and intended level of work without you specifying every step. Continue to state important constraints, approval boundaries, and success criteria explicitly. > Original image detail: GPT-5.6 preserves the…

> Avoid generic brevity instructions That part is confusing because it's not like they provide an example of how default GPT-5.6 output compares with GPT-5.5 both with default output and prompted for brevity. Whenever I use such prompts, it's usually because I want the model to give me the gist in a few sentences. I'd be stunned if GPT-5.6 was that concise by default. I would think that could "break" a lot of things…

For sure verbal diarrhea can be a problem. I think there's a difference between a generic instructions e.g. "be brief" and contextual guidance: "I am an experienced software developer with a recent undergraduate degree in pure mathematics. Be terse, I will ask questions if I need clarification."

Re: GPT-5.6

#760

Post by sidgarimella »

winwang wrote: post #757
sidgarimella wrote: post #756

I've found Sol's propensity for delegating to subagents can make it... disastrously expensive, especially with each subagent having some implicit floor on further reasoning/context gathering before action. The base model is certainly cheaper and more token efficient etc, but on large tasks cost in some way is now n^2

How bad is it for you? Are you on ultra or xhigh/max? I typically ask it (5.5, now 5.6-sol) to use subagents for specific things anyway. On the Pro 20x plan, I'm seeing like ~1% usage per 20-30 min per session (on max effort), which is in line with 5.5. Currently trying out ultra on a personal project, feels like ~3x more expensive per unit time. (No idea on quality yet, for obvious reasons.)

Max via Cursor, should've mentioned, very possible I'm seeing more Cursor than OAI. All the subagents it picked were also Sol Max... I've seen 7-10 in one turn. Regardless the top reasoning tiers being conflated with subagent count feels double edged
Post reply on HN