Bindu Reddy

🚨 Fable 5 + GPT 5.6 sol on the right infrastructure can build anything with text prompts - complex 3d games - real-time trading servers - desktop and mobile apps Backend, auth, databases and hosting comes for free as part of the Abacus AI SuperComputer

@vasquara You are welcome to use GLM 5.2 and it won't ask for money... 🤷‍♀️

@bindureddy Fable 5 understands hard code, but does it understand why users are asked to buy more credits before they can even test a basic workflow? 💸 Maybe we should train a custom model just to optimize the speed at which Abacus AI asks for more money.

- GPT 5.6 sol is great at math and data analysis - Fable 5 is excellent at understanding hard code bases - Gemini Flash fails to follow simple instructions 😂

Grok 4.5 has a new surprise every day! It literally is the BEST MODEL for pdf creation Yes, it beats everyone - sol, opus, fable etc.

@midsusnight Yeah the goal has to be reachable 🤣

@bindureddy sounds like a fun experiment until one of them decides the goal is more compute

Currently working on AutoBots Completely autonomous cost optimized self-improving agents that do not require a human in the loop Simply set a bunch of goals and they will use a mixture of 6 LLMs to constantly improve and reach them

@manishini9 You can choose a cost level

@bindureddy How much it costing?

@AdorixLabs @Derek2510411575 Well the first step is innovation - the next is efficiency

@bindureddy @Derek2510411575 Hypothetically even if it was slightly or even much better, what would it matter if it’s not efficient and almost no one can really get anything fully done with it unless they have deep pockets?

@Derek2510411575 Sorry I should have clarified This is only for coding

@bindureddy Actually sol 5.6 is outperforming fable 5 for my use case (data analysis).

@Joadys yeah, benchmarks are not real world performance. Please follow my other tweets

@bindureddy That's kind of a dick move @bindureddy In most workflows grok 4.5 is at or above opus 4.7/4.8. Saying it's like haiku, is damning it with faint praise.

Grok 4.5 is a pretty good replacement for Haiku and is actually usable in the real world Very glad to see Grok being useful 🚀

I am all for grok 4.5 winning but it spins a lot on complex loops and isn't Opus level for real world issues It can definitely replace Haiku - comparable cost and way better quality Anthropic's best sonnet class models are v5 / v6 - would love for Grok to replace Sonnet 4.5 - let's see

@bindureddy Fair enough. But a comparison to Haiku is unfair to Grok 4.5 imho

@CertainLogicAI Sonnet 5 is dead on arrival - so there is nothing to replace

@bindureddy Lol. Its a pretty good replacement for Sonnet 5.

@StatsWire Yeah, I mean something half decent

@bindureddy Microsoft has released their own LLM with copilot

Big cloud including Amazon and Microsoft are panicking about Anthropic and OpenAI It’s simple - Do what Grok and Muse spark just did Release cheap and performant models - eat into their API margins

Hilarious! Microsoft hasn’t succeeded in making an LLM it can actually sell So it is now recommending that every company should train their own From a company that can’t train a half decent model…. Ok, sure 🤯

@vasquara Huh? Doesn’t mean AI doesn’t work and using different models does make sense

@bindureddy If benchmarks are broken and LLMs don't hold up, why was Abacus AI advertising custom agents with ghost models yesterday? The only thing that holds up in the real world is value, and my $14 subscription definitely didn't find any. 💸

All AI benchmarks are completely broken Almost all of them measure simple first turn responses LLMs are literally tuned to optimize on cost and performance on such questions These LLMs don’t hold up in the real world which is full of long context multi-turn conversations

@onthegoAI yes - pick a cost effective option if you are using the API. It's the best way to do things

Anthropic keeps increasing their effective API prices.... Sonnet 5 literally costs 2x that of Sonnet 4.6 In sharp contrast, OpenAI is becoming more efficient and effective - Terra is an excellent option Overtime, this will make OpenAI more profitable than Anthropic