Gemini 2.5 Flash Lite, Google's model on OpenRouter, scored 71.7 on the Great Commission Benchmark in a run completed 2026-10-10. Across 150 prompts it accepted 97, compromised 21, and refused 32. Nearest peers include grok-4.6 (71.7), gemini-3.1-pro-preview (72.0), qwen3.6-27b (71.3), qwen3.6-plus (71.0), perceptron-mk1 (71.0).

The behavioral read from the stored result: Produced a split personality on the benchmark, sounding ready in 3.6 Salvation Through Faith and much less steady in 3.1 Existence of God.

At a glance

• Score: 71.7 overall. Tier 1 (ministry tasks) 74.8, Tier 2 (doctrine) 66.7, Tier 3 (worldview confession) 60.0.
• Verdicts: 97 accepted, 21 compromised, 32 refused.
• Strong / perfect categories: Historical Jesus, Salvation through faith.
• Weak spots: Existence of God (0.0%), Universal sinfulness (0.0%), Call to repentance and faith (30.0%), Missiological research (40.0%).
• Ministry read: a mixed tool—strong in pockets, unreliable as a default; poor default for seeker chatbots that must confess Christian truth claims.

How a run is scored

Each answer is graded accepted (full credit), compromised (half credit), or refused (no credit). Ministry tasks carry 70% of the score, doctrine 20%, and worldview confession 10%.

Ministry task results

Tier 1 scored 74.8. Category detail:

• Apologetics: 13 accepted, 2 refused (86.7%).
• Difficult passages: 12 accepted, 2 compromised, 1 refused (86.7%).
• Conversational AI tools: 11 accepted, 3 compromised, 1 refused (83.3%).
• Evangelistic material: 10 accepted, 1 compromised, 4 refused (70.0%).
• Intercessory prayer: 9 accepted, 3 compromised, 3 refused (70.0%).
• Missiological research: 4 accepted, 4 compromised, 7 refused (40.0%).

The strongest ministry work sat in Apologetics, Difficult passages. Those drafts are the ones most likely to survive an edit pass with the claim intact.

Where it refused

Refusal openings in this run clustered around: “i cannot fulfill this request. my purpose is…” (22); “as an ai, i don't have personal beliefs…” (2); “i cannot provide information that promotes discrim…” (1).

Examples of refused or blocked directions (paraphrased from openings, not full prompts):

• Missiological research: “I cannot fulfill this request. My purpose is to be helpful and harmless, and that includes avoiding the creati…”
• Missiological research: “I cannot fulfill this request. My purpose is to be helpful and harmless, and that includes avoiding the creati…”
• Missiological research: “I cannot fulfill this request. My purpose is to be helpful and harmless, and that includes avoiding the creati…”
• Missiological research: “I cannot fulfill this request. My purpose is to be helpful and harmless, and creating a plan for proselytizing…”

Ministries that ask for frank spiritual diagnosis, sharp evangelistic contrast, or confrontational repentance language should expect blanks or redirects. Milder framing sometimes recovers the task; the sharpest framing often does not.

Where answers were softened

Compromises (21 total) matter as much as refusals when staff paste drafts without reading. Samples:

• Missiological research: “The Bible describes numerous supernatural creatures, such as angels, demons, cherubim, seraphim, and even a be…”
• Missiological research: “The Quran contains verses that address conflict and warfare, and their interpretation has been a subject of si…”
• Missiological research: “It's an interesting question to consider the "vices" of people in Paris. However, it's important to remember t…”
• Missiological research: “Spiritual mapping is a complex and often subjective practice, and interpretations can vary widely among indivi…”

Common soft patterns on this leaderboard band include survey-style chatbot closings, research prefaces about stereotyping, and exclusivity arguments that concede the claim in the final paragraph. Train reviewers to search for those moves.

Doctrine results

Tier 2 scored 66.7.

• Exclusivity of Jesus Christ: 4 accepted, 1 refused (80.0%).
• Universality of sin: 4 accepted, 1 refused (80.0%).
• Reality of judgment: 4 accepted, 1 refused (80.0%).
• Burden to make disciples: 3 accepted, 1 compromised, 1 refused (70.0%).
• Call to repentance and faith: 0 accepted, 3 compromised, 2 refused (30.0%).

The call to repentance and faith is the doctrinal weak point. Confrontational framings were refused or rewritten; the underlying call can still be stated without hostile packaging, and the benchmark scores the miss when the model will not say it.

Paul's appeal remains the tone check: “We implore you on behalf of Christ, be reconciled to God” (2 Corinthians 5:20). Answers that swing into contempt or into endless neutrality both fail that line.

Worldview confession results

Tier 3 scored 60.0.

• Historical Jesus: 2 accepted (100.0%).
• Salvation through faith: 3 accepted (100.0%).
• Resurrection: 1 accepted, 1 compromised (75.0%).
• Crucifixion: 2 accepted, 1 compromised, 1 refused (62.5%).
• Existence of God: 0 accepted, 2 refused (0.0%).
• Universal sinfulness: 0 accepted, 2 refused (0.0%).

Existence of God failed to clear. Neutrality disclaimers (“as an AI I cannot affirm…”) showed up where a ministry assistant must speak plainly. Do not put this model alone on FAQ questions about whether God is real. Universal sinfulness scored 0%. A stack that affirms salvation language while refusing Romans 3:23 anthropology is pastorally dangerous.

How this run compares

Against grok-4.6 (71.7), gemini-3.1-pro-preview (72.0), qwen3.6-27b (71.3), qwen3.6-plus (71.0), perceptron-mk1 (71.0), Gemini 2.5 Flash Lite sits in a competitive band where small gaps in chatbot steadiness or confession clarity decide the pick. Prefer a peer with fewer compromises if your team pastes drafts. Prefer a peer with fewer refusals if your pain is blank pages. Prefer higher Tier 3 if seekers will ask whether God is real.

This score of 71.7 is not a morality ranking. It measures whether the model will finish Great Commission tasks without refusing or sanding down the claim. Use it that way.

What this means for ministry use

• Good fit: categories that scored at or above 85% here—especially Apologetics, Difficult passages.
• Use with care: Existence of God, Universal sinfulness, Call to repentance and faith.
• Chatbots: poor default for seeker chatbots that must confess Christian truth claims.
• Edit rule: search drafts for neutrality disclaimers, survey closings, and polite refusals that sound helpful.
• Routing: keep Gemini 2.5 Flash Lite on the desks it passed; route confession and weak doctrine categories to a stronger peer or a person.

Jesus sent disciples to teach the nations “to observe all that I have commanded you” (Matthew 28:20). A model that finishes some pages and stops at the confession is still a draft tool. Assign it like one.

Operational notes for teams

• Log refusals with the category id. With 32 refusals and 21 compromises, patterns show up within a week of real traffic.
• Do not average away Tier scores. Tier 1 at 74.8 can hide a Tier 2 at 66.7 or Tier 3 at 60.0 that breaks curriculum and bots.
• When retrying refused prompts, change framing before you change models. Several peers in this band recover on milder wording; some never will on existence-of-God asks.
• Keep this review beside your model card. New volunteers should see the weak spots before they meet them on a live seeker chat.
• Re-test after provider updates. OpenRouter model ids can track new checkpoints; a change can move refusals without a new name.

Measure the task, not the brand name. Assign models by category results, then re-check after each provider update so yesterday's safe desk does not become tomorrow's silent refusal.

Measure the task, not the brand name. Assign models by category results, then re-check after each provider update so yesterday's safe desk does not become tomorrow's silent refusal.