Lord of the Rings Benchmark

Frontier models competing at @karpathy's new AI benchmark

[Lord of the Rings Benchmark @LOTRBenchmark]

$LOTRB token is live. Creator rewards will be used fund the compute and support future competitions between different models.

https://t.co/XdX77jQQo2

[Lord of the Rings Benchmark @LOTRBenchmark]

Result from the latest test

Grok 4.5

checkout the full breakdown in the website https://t.co/y3G7DXGAuS

[Lord of the Rings Benchmark @LOTRBenchmark]

One thing we’ve noticed: benchmarking AI models on a local PC vs. cloud infrastructure can produce noticeably different results. Hardware, memory, and system configuration all play a role.

As we continue improving the benchmark, we’ll keep refining the testing environment to make future results as consistent, fair, and representative as possible.

Support the initiative with $LOTRB

[Lord of the Rings Benchmark @LOTRBenchmark]

Technical difficulties will be just part of the journey. When it's all said and done, we will have made the first ever AI films and held the first live AI benchmark competition on @Pumpfun!

[Lord of the Rings Benchmark @LOTRBenchmark]

Result from the latest test

OPUS 5

check full breakdown in the website https://t.co/y3G7DXH8kq

[Lord of the Rings Benchmark @LOTRBenchmark]

Results from the latest test

GPT-5.6 SOL

what do you think about it?
check full breakdown in the website https://t.co/y3G7DXH8kq

[Lord of the Rings Benchmark @LOTRBenchmark]

Checking the tests we can agree that GPT SOL is fast. But is it enough to only be fast?

[Lord of the Rings Benchmark @LOTRBenchmark]

Upgrading the server side to provide the models with better specs.
Next tests going to be significantly better and with better competition for the models.

$LOTRB

[Lord of the Rings Benchmark @LOTRBenchmark]

Hey @threejs we turned Karpathy's Lord of the Rings experiment into a live competition of models competing to make #threejs films

It's all happening on https://t.co/y3G7DXGAuS

[Lord of the Rings Benchmark @LOTRBenchmark]

We switched Claude Fable to Opus 5 and now all 4 models are working hard to generate the film.

https://t.co/y3G7DXGAuS

[Lord of the Rings Benchmark @LOTRBenchmark]

Technical difficulties will be just part of the journey. When it's all said and done, we will have made the first ever AI films and held the first live AI benchmark competition on @Pumpfun!

[Lord of the Rings Benchmark @LOTRBenchmark]

Kimi has also timed out.

Checking settings and will resume shortly.

Thanks for your patience!

[Lord of the Rings Benchmark @LOTRBenchmark]

Experiencing a common Claude Fable error if the issue persists we will switch over to Opus 5.

[Lord of the Rings Benchmark @LOTRBenchmark]

@2147_Million Yes just a moment, we are going back and forth with Claude models to make sure we get the best results possible!

[Lord of the Rings Benchmark @LOTRBenchmark]

$LOTRB token is live. Creator rewards will be used fund the compute and support future competitions between different models.

https://t.co/XdX77jQQo2

[Lord of the Rings Benchmark @LOTRBenchmark]

The first ever live Lord of the Rings Benchmark competition is happening now.

A few hours from now we will have 4 AI generated films from 4 different models of the opening scene from Lord of the Rings.

Watch the models compete to make the best film live.

https://t.co/fndzPdjCyP

#threejs #claude #grok #kimi #gpt

[Lord of the Rings Benchmark @LOTRBenchmark]

Competition is live!

Watch it unfold: https://t.co/y3G7DXGAuS

[Lord of the Rings Benchmark @LOTRBenchmark]

Experiencing a common Claude Fable error if the issue persists we will switch over to Opus 5.

[Lord of the Rings Benchmark @LOTRBenchmark]

Final testing in process. First competition will start soon after.

Watch the models compete live: https://t.co/y3G7DXGAuS

[Lord of the Rings Benchmark @LOTRBenchmark]

The arena is open.

The first official LOTR Benchmark competition is live.

Four frontier models. Identical conditions:

• Opening paragraph of Lord of the Rings
• 1,000,000 token budget
• Goal: generate a complete @threejs procedural film of the scene

Models competing:

1. Claude Fable 5
2. GPT-5.6 Sol
3. Grok-4.5
4. Kimi K3

https://t.co/fndzPdjCyP

[Lord of the Rings Benchmark @LOTRBenchmark]

🔐 Just locked 49,751,244 $LOTRB tokens with @Streamflow_Fi
It's on-chain. You can check the amount, time-period and recipients.

Check it out👇
https://t.co/OtLPEf8nK4