ikindatried

@Borg_Cryptos 4-3

the plan is first build a very very strong infra so you can give it for free for sometime and then, start to charge. a very lame and bulletproof plan you know.

@elonmusk @minchoi how can kimi be ahead of you guys, it really don't make sense.

@minchoi Our 2T model, which is better than our 1.5T in every way, will finish initial training next week. It might be able to exceed Kimi, but with speed and token efficiency close to our 1.5T (aka Grok 4.5).

Grok 4.5 is only 1.5T parameters. Kimi K3 is 2.8T... and costs 3× more per task. SpaceXAI cracked intelligence efficiency. Now imagine Grok at 3T.

@CommandCodeAI am i the only one that likes sol here.

Kimi K3 vs GPT-5.6 Sol vs Fable 5 Three top-tier models. Same /design prompt across all. Parameters reviewed: Pace, design sense, gameplay feel Result: Kimi K3 nails the design sense and knows exactly what to add in-game. Pace and gameplay held up well. Fable 5 and GPT-5.6 Sol both play too fast, video isn't sped up. Ranking (DX, Features & Cost): → Kimi K3: 9.5/10 · $0.030 → Fable 5: 7.5/10 · $0.38 → GPT-5.6 Sol: 7/10 · $0.11

@thsottiaux Sol it is then

How do you pronounce Sol

there should be donate reset option too.

i am absolutely flexing this.

@Elaina43114880 i am absolutely loving this. got $100 credits btw.

OpenAI and Anthropic just reset everyone’s limits almost at the same time! 😻 Codex and ChatGPT Work users are getting their full weekly usage back, while Claude reset both the 5-hour and weekly limits for all users. As a user, I’m very happy to watch OpenAI and Anthropic keep taking shots at each other, especially if every round comes with another usage reset.... 😹👀

$100 in credits letsss goooo. thanks to our lord @thsottiaux @ajambrosino

@strimblez You had it you know.

if you had 5 days of unlimited free tokens for codex, what would you build?

@rezoundous It's like you don't have to even specify that. It just gets shit done.

Every prompt on GPT-5.6 is like a goal now

GPT-5.6 Sol is sharp at long-horizon tasks. Gave it a goal to add proper observability to a small internal service. It spun up subagents to map the codebase, found every relevant logging point across 9 files, implemented structured tracing with context propagation, wrote the config, added basic tests, and even suggested a simple query for the dashboard. Kept everything consistent without me stepping in. Really useful when you have massive tasks. Ability to handle a lot of tasks and proper use of tools and skills that and more alignment with the same skill ,for example if you use image generation from it, it is more better at utlising that, and that too for long time reliably is what 5.6 sol is.🐐

@OfficialLoganK fix Gemini for agentic coding.

it’s surprising to me how many people seem to not understand that great models are built with super high quality curated data finding novel ways to create / get this data is a huge edge

@tunguz sometimes it is buggy.

Oh wow, ALL of my OpenAI usage limits got reset *RIGHT NOW*!

@bindureddy except for ui.

@Yuchenj_UW openai itself needs a model router here.

We desperately need a smart model router. 1. We’re seeing a model explosion: GPT-5.6, Grok 4.5, Muse Spark 1.1, GLM-5.2, and Fable 5 all launched within the past month. 2. Even for a single model family like GPT-5.6, there're 3 (Sol, Terra, Luna) and 5 reasoning-effort levels. That is far too many decisions for users to make manually. The best model should be selected automatically based on the task, latency, quality, and cost.

@MatthewBerman absolutely make sense though. google is beyond cooked here. it seems way way behind. they are absolutely not getting it.

Mark my words: Codex / Claude Code will own the browser market within 12 months.

@weswinder I saw this a bit too late.

hot take: you should basically never use ultra effort with 5.6 sol in codex > it uses WAY too many tokens > your usage will get drained super fast (even on pro) > codex usage tracker doesn't update fast enough for you to react until it's too late learned this the hard way

@argofowl Made a rookie mistake to use sol extra high.

gpt 5.6 sol medium and high (sometimes) is the way to go