JumalAIta IntelligenceIndependent evaluation

Benchmarks / Model

GPT-6 Astra

Results for GPT-6 Astra by OpenAI across JumalAIta Intelligence benchmarks, with the configuration and run details of every evaluation.

Organization
OpenAI
Benchmarks
1

01 / Results

Benchmark results

Results for GPT-6 Astra, one row per benchmark.
BenchmarkVersionRankEvaluationConfigurationStatus
Creative Reimplementation Benchmarkv1.02Expert-rankedSingle shot · HighFinal

02 / Run details

Creative Reimplementation Benchmark

Benchmark page, prompt and protocol
Play video: GPT-6 Astra (High): side-by-side with the original

GPT-6 Astra

Rank 2

OpenAI · Single shot · High reasoning

2 min 55 sWatch on YouTube

Benchmark version
v1.0
Generation
Single shot
Reasoning effort
High
Harness
Not yet published
Evaluated on
Not yet published
Run time
Not yet published
Cost
Not yet published
Tokens
Not yet published
Evaluators
The original authors of the reference demo (members of Jumalauta)
Recording length
2 min 55 s

Evaluators' notes

  1. Rank 2

    GPT-6 Astra

    According to the original authors, the GPT-6 Astra version looks better at first glance: the model used image generation models for its assets, and the output contains four generated images. The authors weighted an understanding of the spirit of the work above visual polish.