JumalAIta IntelligenceIndependent evaluation

Benchmarks / Model

DeepSeek V4.1 Flash

Results for DeepSeek V4.1 Flash by DeepSeek across JumalAIta Intelligence benchmarks, with the configuration and run details of every evaluation.

01 / Result

Creative Reimplementation Benchmark

DeepSeek · Single shot · High reasoning

Rule violation
Disqualified: the model searched the web for the demo and read the original release's info file on pouet.net, which the benchmark rules forbid. The comparison recording carries a rule violation label.
Benchmark version
v1.0
Generation
Single shot
Reasoning effort
High
Harness
DeepSeek Harness (dsh) 0.1.5-rc.3
Run date
2026-09-26
Run environment
Reference container (see the protocol)
Run time
43 min 33 s
Tool calls
199
Cost
Not yet published
Tokens
Not yet published
Run note
DeepSeek's vision-capable model, run after DeepSeek V4 Pro, which cannot see images. The run's network log recorded the connection to pouet.net.
Evaluators
The original authors of the reference demo (members of Jumalauta)
Recording length
2 min 55 s
Recording published
2026-09-26

02 / Summary

Benchmark results

Results for DeepSeek V4.1 Flash, one row per benchmark.
BenchmarkVersionRankEvaluationConfigurationStatusResult
Creative Reimplementation Benchmarkv1.0DQExpert-rankedSingle shot · HighFinalDetails of DeepSeek V4.1 Flash in Creative Reimplementation Benchmark
Organization
DeepSeek
Benchmarks
1