Joined April 2025
- Tweets 255
- Following 14
- Followers 1,692
- Likes 83
39 Photos and videos
Pinned Tweet
May 7
The most advanced benchmark and sandbox for AI image and video.
Now available at FilmArena.ai
Apr 22
Today we're launching FilmArena [dot] ai.
Biggest release we've ever shipped.
The problem: a new SOTA AI image & video model drops every week. @OpenAI shipped GPT Image 2 literally yesterday. 1000 endpoints, no honest answer to "which one is best right now?"
Arena AI and Artificial Analysis cracked LLM benchmarking. But for creative AI, they're solving the wrong problem.
There is no "best image-to-video model."
Best at what? World building? Action Sequences? Expressions? Every category has a different winner, and it flips weekly.
So we benchmarked 40 SOTA models with 2M votes across 190 countries. By subcategory. Re-ranked every week as new models ship. With thousands of real humans.
But that's just half of it.
Then there's the SANDBOX.
Higgsfield and other aggregators put models in one place. But creators still juggle twelve subscriptions to do one job.
FilmArena collapses it into one tab. Every frontier model. One interface. Same price. Call the best, or call 3 at once and compare.
The AI film stack has been fragmented since day one. Today it isn't.
Go try it free: filmarena [dot] ai
Already wild results from this week:
1. @OpenAI GPT Image 2 (released yesterday) dominates image editing, but @GoogleDeepMind Nano Banana 2 still wins animated
2. @PixVerse_ C-1 beats @dreamina_ai Seedance 2 on animated
3. Veo 3.1 Fast outperforms Veo 3.1 on realistic video
And a lot more. Go check it out.
1
1,979
koyal.ai (YC F25) retweeted
How many of you would be into this limited edition drop of @KoyalAI OKO VIDEO merch?
1
1
3
329
koyal.ai (YC F25) retweeted
who needs a gf when you can t-pose with the homies
Single until Series B -
out this summer starring @contextconor and me.
8
2
21
2,976
koyal.ai (YC F25) retweeted
Jun 12
2
1
10
862
koyal.ai (YC F25) retweeted
Jun 13
This is exciting.
An Indian video model for Indian audiences at Indian prices.
It’s a finetune today, hoping to see a from scratch model tomorrow.
Going to support this model soon for our Indian clients at @KoyalAI
Also may be high time to have cultural benchmarks on filmarena.ai đź‘€
Jun 12
Introducing Varya.
Built by @Avataar, #Varya delivers frontier-quality results at materially lower inference cost.
Deeply grateful to #IndiaAIMission and @SecretaryMEITY for the support, mentorship and conviction in helping bring this to life.
Thank you @RajanAnandan, @brettsingh,
@shaileshlakhani for your invaluable partnership through this journey.
@GoI_MeitY @peakxvpartners
Try it: varya.avataar.ai
3
1
15
3,213
koyal.ai (YC F25) retweeted
Jun 11
at the koyal x meta AI Media event
8
13
1,372
koyal.ai (YC F25) retweeted
Jun 10
Claude Fable is the most creative model in the world.
This is the most capable model for movie script breakdowns, shot lists, camera blocking and video prompting.
You often hear about models like this and their capabilities in coding or finance, but this is the best model for filmmaking.
@KoyalAI now supports it day zero.
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.
Its capabilities exceed those of any model we’ve ever made generally available.
1
1
3
670
Jun 6
đź‘€
KOOKOO
1
616
koyal.ai (YC F25) retweeted
Jun 5
1
1
2
807
Jun 4
Poker nights at Koyal HQ
1
311
koyal.ai (YC F25) retweeted
Jun 3
I'm a big big fan of open source models in Creative AI.
It allows non big-tech labs to shine and build goodwill.
But I'm super confused by this new set of models by @ideogram_ai (already open source) and @krea_ai (soon open source)
Why are these models purely text to image and don't natively support image to image?
I feel with Flux Kontext, everyone took a step forward with models to understand that professional workflows need image to image models, yet this feels a step backwards.
Second, would love for image models to start benchmarking early on Filmarena.ai. We have tie-ups with all the major video labs, but I think a top image model needs to shine in the best benchmarking suite run by @KoyalAI :)
4
1
15
2,378
Jun 3
This is fake news. We have GPUs at Koyal
KOOKOO
3
231
koyal.ai (YC F25) retweeted
Jun 2
A lot of SF companies struggle to have women.
I'm proud to say that we have more women than men at @KoyalAI
It's not that hard honestly.
It starts with having your sister as the CTO :)
Come work with us, gender independent. We're looking for an ML Eng!
3
1
21
2,235
Jun 2
TEAM PIC
- KOOKOO
2
222
koyal.ai (YC F25) retweeted
May 27
I'll give 300 bucks to anyone who can submit a @KoyalAI video to win this challenge
May 27
Lately everyone’s been talking about AI agents, so we got curious and asked our community to share their best “agentic” creative workflows.
We picked 3 to test through @LumaLabsAI, @runwayml and @UtopaiStudios and the big takeaway from all these tests is that agentic creative workflows still don’t seem that impressive in their current form.
So here’s our next challenge: We’ll give $200 to anyone in the comments who can share an ACTUAL agentic creative AI workflow that’s genuinely better than just prompting manually yourself.
1
4
921
koyal.ai (YC F25) retweeted
May 3
We’ve been trained to think that the difference between a good and bad AI video is the list of keywords at the end of a prompt. @meh_agarwal argues that this is fundamentally against human nature and creative intuition.
The future of AI video isn't about learning secret codes; it's about agentic workflows. You describe what you want to see, and the AI agent translates that into the technical language the model needs.
We are moving past the "Infinite Text Box" phase of AI. The next generation of tools won't have prompting guides, they’ll have creative controls.
@KoyalAI
1
1
1
1,447
koyal.ai (YC F25) retweeted
May 4
The "Model Wars" are officially here, and every two weeks, a new leader emerges. @meh_agarwal (@KoyalAI) explains why committing to just one foundational model is a mistake for any serious filmmaker.
Koyal’s "Film Arena" benchmarks every major model: Sora, Runway, Luma, on specific cinematic tasks. Whether it’s slow-mo, action sequences, or micro-motions, there is no single "best" model, only the best model for that scene.
We are moving into a "Model Agnostic" future. Don't limit your creativity to the constraints of one lab when you can use an agent to cherry-pick the best pixels in the world.
1
1
1
817
koyal.ai (YC F25) retweeted
May 7
Share the craziest stuff you’ve made using FilmArena.ai
I’ll send you some @KoyalAI merch
1
1
8
443
koyal.ai (YC F25) retweeted
Apr 28
Last year, I thought to myself, it would be so cool to work with athletes once we launch. Had no idea how to actually make it happen… and tbh I still don’t.
Just trust serendipity.
- @dan4kmp4 and team are great at filmmaking. Definitely consider them for your next launch video.
- special thanks to @KoyalAI x @MultifactorCOM for letting us use their amazing office space :D
Apr 28
Introducing Maverick #1
Max "Pain" Griffin, UFC fighter
- started using Mave headset 100 days ago
- for stress regulation and performance
Brain score: 71 (day 0) -> 83.4 (day 30)
Grab your headset: mavehealth.com
2
4
25
2,695
koyal.ai (YC F25) retweeted
Apr 30
When ChatGPT first dropped, the tech world was thrown into a tailspin, and the "experts" were suddenly playing catch-up. @meh_agarwal shares a story of realization: you don’t need a 20-year career to be the authority in the room.
If you are studying LLMs and machine learning today, you are the frontier. While the veterans are trying to figure out how to pivot, the students and young founders are already building the future.
Don't let age intimidate you when the technology itself is only a few years old. In this race, the person with the most recent knowledge takes the lead.
@KoyalAI
1
1
5
1,066