Hashim Syed

Co-founder of @straitlyai and @xynth_m | Agentic ai @Meta, Infra @awscloud | @UBC

[Hashim Syed @Hashim__Syed]

@ReddyYellender @omarsar0 and a refactor that touches a bunch of files goes straight to Opus 5.5, so it actually stays consistent

[Hashim Syed @Hashim__Syed]

@joshtheaiguy @omarsar0 that's the whole idea. and if you ever want a say, tell Jev in plain English on the JevRouter page, like "use Astra for backend", and it sticks

[Hashim Syed @Hashim__Syed]

@King_of_Calera @omarsar0 Jevs only read, they never write. each one scores its own slice, and the pick happens once after the votes are in, so there's no shared state to race on. then one model runs the task at a time, so nothing races on your code either

[Hashim Syed @Hashim__Syed]

@TheManuAgrawal @omarsar0 thank you!

[Hashim Syed @Hashim__Syed]

@omarsar0 this is exactly what we built it for. one session, the right model on each part, and you never touch the picker. thanks for putting it through a real build

[Hashim Syed @Hashim__Syed]

@shiri_shh @straitlyai yeah most auto routers fail at this

[Hashim Syed @Hashim__Syed]

@omarsar0 @straitlyai I would say its really good for agentic tasks as you can see.

And also pinning providers with just a prompt is pretty cool too.

[Hashim Syed @Hashim__Syed]

@againoa1 @straitlyai Would love to check it out if you have a site!

[Hashim Syed @Hashim__Syed]

@Av1dlive @straitlyai Yeah, but it has to be done right. The next version will be fine-tuned with 256k context to reduce parallelization.
I really think this is the future of LLMs: a mixture of models, routed on demand for specific use cases.

[Hashim Syed @Hashim__Syed]

@alwayspriyesh @straitlyai This looks very interesting!

[Hashim Syed @Hashim__Syed]

@PromptLLM The future of llms, on demand mixture of models suited best for your workloads.

[Hashim Syed @Hashim__Syed]

@Im_IrushiK Very interesting use case! Looks like one shotting apps without having to worry about costs is possible now.

[Hashim Syed @Hashim__Syed]

@shiri_shh Hermes agent is basically free with jevRouter

[Hashim Syed @Hashim__Syed]

@powl_d 30% off opus 5.5 !

[Hashim Syed @Hashim__Syed]

@alwayspriyesh JevRouter!

[Straitly @straitlyai]

This is JevRouter.

99% of Opus 5.5 and GPT-6 Astra's intelligence, for 40% of the cost.

Hundreds of Jevs read your ENTIRE request and analyze every single benchmark across all models to match you with the perfect model.

Enjoy 30% off all frontier models

[Hashim Syed @Hashim__Syed]

@okkshitij how do i use this

[Mujtaba @Mutchtaba2]

1:00AM launch shoot hits different. Only in SF @frontiertower

[Hank Couture @HankCouture]

Glossed over in every success story is a period of wandering curiously into the unknown

You must make time to wander

That is where all great stories begin

[Hashim Syed @Hashim__Syed]

@CodeAi_Vishal @Mutchtaba2 Thank you !