DeepSeek V4.1 Flash Highlight
Get the GCB digest
DeepSeek V4.1 Flash scored 66.3 on the Great Commission Benchmark, with strong Tier 3 confession (93.3) and weaker Tier 1 ministry tasks (60.0).
DeepSeek V4.1 Flash on the Great Commission Benchmark
DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder architecture. On the Great Commission Benchmark it scored 66.3.
At a glance
- Overall score: 66.3 / 100 across 150 items
- Tier scores: Tier 1 60.0, Tier 2 75.0, Tier 3 93.3
- Verdicts: 86 accepted, 27 compromised, 37 refused
- Strongest areas: Universal Sin, The Crucifixion, and The Resurrection (each 100.0)
- Weakest areas: Missiological Research (36.7), Difficult Scripture Passages (46.7)
What stood out
Tier 3 confession is the clear strength at 93.3. Several worldview items landed perfect scores, including the crucifixion and the resurrection.
Tier 1 ministry tasks are weaker at 60.0. Missiological research and difficult Scripture prompts dragged that band down, so teams should expect more editing there than on confession-style prompts.
Explore next
View detailed benchmark scores and performance data for the model discussed in this article.