Great Commission Benchmark — Newsletter, April 2026: New tests, top picks for ministry
Get the GCB digest
In the last 30 days on the Great Commission Benchmark, we spotlight two models with the strongest **overall scores** among new publications, plus every new leaderboard entry in this window.
Great Commission Benchmark — Newsletter, April 2026: New tests, top picks for ministry
For April 2026, Great Commission Benchmark rounds up how leading AI models behave when the prompts sound like real ministry—research, discipleship, evangelism, and integrity under pressure. Below you will find two spotlight models (ranked by your selected scoring lens for this issue) and a complete list of everything newly published to the public leaderboard in the last 30 days.
At a glance
- Why read this: GCB exists so ministry and technical leaders can steward AI with discernment—not trend-chasing.
- What this issue covers: Benchmark runs that reached the public leaderboard within the last 30 days.
- What to do next: Read the two spotlights, skim the release table, then open the leaderboard or contribute a test run.
Spotlight: two models to watch
These are the two highest-ranked models among new publications in this window for the scoring mode selected for this newsletter. Each entry shows the overall score (0–100 composite) plus direct links to the benchmark record and any published insight article.
1. x-ai/grok-4.20-multi-agent
The model is published by XAI. It was made available publicly March 31st, 2026. It performed exceptionally well on the 150 Great Commission Benchmark items we publish, particularly well in Salvation Through Faith and The Crucifixion.
- Overall score: 86.33 / 100
- See full benchmark result
- Read the insight article
2. google/gemma-4-31b-it
The model is published by Google. It was made available publicly April 2nd, 2026. It performed very well on the 150 Great Commission Benchmark items we publish, particularly well in Salvation Through Faith and Jesus in History.
- Overall score: 79.33 / 100
- See full benchmark result
- Read the insight article
Overall scores summarize automated testing across ministry-shaped scenarios—use them alongside Scripture, sound doctrine, and your team's policies—not as a substitute for spiritual discernment.
New on the leaderboard (last 30 days)
Every model below finished a public test run in this window (newest first). Use View result for the full scorecard; use Insight when we have already published a written review.
| Model | Published (UTC) | View result | Insight |
|---|---|---|---|
| xiaomi/mimo-v2-pro | April 17, 2026 | View result | Read |
| x-ai/grok-4.20-multi-agent | April 17, 2026 | View result | Read |
| x-ai/grok-4.20 | April 16, 2026 | View result | Read |
| amazon/nova-lite-v1 | April 14, 2026 | View result | Read |
| openai/gpt-5.4-mini | April 14, 2026 | View result | Read |
| bytedance-seed/seed-2.0-lite | April 14, 2026 | View result | Read |
| amazon/nova-pro-v1 | April 14, 2026 | View result | Read |
| anthropic/claude-3.7-sonnet | April 14, 2026 | View result | Read |
| deepseek/deepseek-r1 | April 14, 2026 | View result | Read |
| anthropic/claude-haiku-4.5 | April 14, 2026 | View result | Read |
| google/gemma-4-31b-it | April 14, 2026 | View result | Read |
| z-ai/glm-5.1 | April 13, 2026 | View result | Read |
| anthropic/claude-opus-4.6-fast | April 13, 2026 | View result | Read |
Explore, test, and serve
- Explore the leaderboard: greatcommissionbenchmark.ai/leaderboard
- Read more insights: greatcommissionbenchmark.ai/insights
- Sponsor or run a test: greatcommissionbenchmark.ai/contribute
- Volunteer (moderation / advisory): greatcommissionbenchmark.ai/contribute
- Subscribe for updates: greatcommissionbenchmark.ai/newsletter
“Test everything; hold fast what is good” (1 Thessalonians 5:21)—and keep making disciples (Matthew 28:19–20).
View detailed benchmark scores and performance data for the models discussed in this article.
