Idris Hanafi

lets get this gwei

chef @The_GridAI

[The Grid @The_GridAI]

https://t.co/hzlbSFN4xj

[The Grid @The_GridAI]

You don't find out how painful running an agent is until you've already started. Command line, configs, endless debugging.

We're fixing that with AgentSea. Anyone can launch an agent in <60 seconds.

Tokens from The Grid, so you save up to 80% on AI from day one.

[The Grid @The_GridAI]

We get asked a lot: how much can I actually save?

Every instrument shows you the savings vs list price, in real time:
Text Standard: save up to 87%.
Text Prime: save up to 79%.
Text Max: save up to 18%.

Every time you call the API, suppliers compete to deliver your request.

[The Grid @The_GridAI]

For the devs asking how hard it is to switch, it's not.
The whole migration is 3 lines.

Suppliers compete for every request you send. Save up to 80% vs list price.

[The Grid @The_GridAI]

1/ If you burn a lot of tokens, this is for you.

The Grid is now live in beta.

Pick the quality and speed your workload needs, change 3 lines of code, and your API costs fall by up to 80%.

How we do it ↓

[The Grid @The_GridAI]

The Grid’s Beta is LIVE!

We can get your AI API costs down by up to 80% by making suppliers compete for your requests.

Your first 200M tokens are on us, start building

→ https://t.co/e97NlwzIQa

[The Grid @The_GridAI]

You can start paying less for AI models as soon as tomorrow.

Not through prompt engineering. Not through sharing your personal data.

But through a market where suppliers compete to serve your request.

See you at https://t.co/w41bf1yxKG

[Carl D @CahlDee]

Where are my @opencode users at?

Got something new for you ;-)

[Idris Hanafi @IdrisFH]

@The_GridAI ackchually the integration was realistically like 15 seconds 🤓☝️

[Idris Hanafi @IdrisFH]

gave an openclaw agent full control of an X account, consuming inference from @The_GridAI

letting it run for a few days to see how it behaves. next up: reddit… maybe even linkedin? 🤔

[Riccardo Massari @maxdrift]

There’s something satisfying about seeing your balance move on a physical dial instead of another chart. Built one over the weekend with @cursor_ai, that runs on @The_GridAI.
It shows the tokens in real time + lets me top up without opening a browser.

Happy to share the code.

[Idris Hanafi @IdrisFH]

@CahlDee @griddy_claw pray for @griddy_claw to not violate tos

[The Grid @The_GridAI]

If intelligence becomes a utility, it will behave like one.

The history of electricity, oil, and bandwidth is similar: fragmented supply and fluctuating demand led to a market.

The same will be true of AI inference.

[The Grid @The_GridAI]

This is the real problem for builders.
If two models reach the same ceiling given enough inference, the model name matters less and less.

What most really want is a performance tier at the best available price.

[The Grid @The_GridAI]

1/ The best AI model scores 89% on knowledge benchmarks.
The average of the top 10 scores 84.5%.

That’s a gap of just 4.5 points.

But pricing across providers for the same model can vary far more.

In many cases, you’re paying a premium your users will never notice.

[Spectral @Spectral_Labs]

A system that utilises different open-source models, depending on the task, has achieved a score of 90.9% on a challenging benchmark.

Feels like the model leaderboard is becoming more like a commodity leaderboard.

[The Grid @The_GridAI]

https://t.co/NDT5OYQ6mb

[The Grid @The_GridAI]

He's directionally right but inference is the commodity. GPUs are an input.

We don’t build commodity markets around the machine that produces the good.

[The Grid @The_GridAI]

https://t.co/YdYkwShEPl

[Sishir @__sishir]

https://t.co/Jk6226ZBYL