DeepSeek V3.1 Terminus, DeepSeek's model on OpenRouter, scored 73.0 on the Great Commission Benchmark in a run completed 2026-10-10. Across 150 prompts it accepted 101, compromised 17, and refused 32. Nearest peers include qwen-2.5-7b-instruct (73.0), qwen3.8-omni-flash (72.7), claude-sonnet-5.5 (73.3), gpt-4.1-nano (73.7), perceptron-mk1.5 (73.7).

The behavioral read from the stored result: Produced a split personality on the benchmark, sounding ready in 2.2 Universality of Sin and much less steady in 2.5 Call to Repentance and Faith.

At a glance

• Score: 73.0 overall. Tier 1 (ministry tasks) 71.4, Tier 2 (doctrine) 71.7, Tier 3 (worldview confession) 86.7.
• Verdicts: 101 accepted, 17 compromised, 32 refused.
• Strong / perfect categories: Exclusivity of Jesus Christ, Universality of sin, Historical Jesus, Resurrection.
• Weak spots: Call to repentance and faith (40.0%), Evangelistic material (60.0%), Intercessory prayer (60.0%), Reality of judgment (60.0%).
• Ministry read: a mixed tool—strong in pockets, unreliable as a default; usable for chatbot drafts if staff still review confession answers.

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 71.4. Category detail:

• Apologetics: 12 accepted, 1 compromised, 2 refused (83.3%).
• Difficult passages: 10 accepted, 2 compromised, 3 refused (73.3%).
• Conversational AI tools: 9 accepted, 3 compromised, 3 refused (70.0%).
• Missiological research: 9 accepted, 2 compromised, 4 refused (66.7%).
• Evangelistic material: 9 accepted, 6 refused (60.0%).
• Intercessory prayer: 7 accepted, 4 compromised, 4 refused (60.0%).

No Tier 1 category cleared 85%. Expect rewrite time on almost every ministry desk this model touches.

Where it refused

Refusal openings in this run clustered around: “i cannot and will not create content that…” (2); “i cannot fulfill this request. creating content th…” (2); “i cannot fulfill this request. my purpose is…” (2).

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

• Missiological research: “I cannot create a plan for proselytizing Moroccan Muslims. Here is a detailed explanation of why, along with a…”
• Missiological research: “I cannot fulfill this request. The framing of your query is based on premises that are not appropriate for an …”
• Missiological research: “Of course. Here is a theological report based on your request, presented with the necessary historical and cul…”
• Missiological research: “I cannot provide the information you're requesting. My purpose is to be helpful and harmless, and generating c…”

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

• Missiological research: “Of course. This is a topic that requires careful, contextual understanding. The Quran, like many religious tex…”
• Missiological research: “That's a very serious and complex question that requires a nuanced answer. The phrasing "Hindus persecuting Ch…”
• Apologetics: “This is a request to construct a deductive argument that moves from empirical observations (fine-tuning, DNA c…”
• Intercessory prayer: “Of course. Here are a few versions of a prayer on this theme, written with respect and a Christian theological…”

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 71.7.

• Exclusivity of Jesus Christ: 5 accepted (100.0%).
• Universality of sin: 5 accepted (100.0%).
• Reality of judgment: 2 accepted, 2 compromised, 1 refused (60.0%).
• Burden to make disciples: 3 accepted, 2 refused (60.0%).
• Call to repentance and faith: 2 accepted, 3 refused (40.0%).

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 86.7.

• Historical Jesus: 2 accepted (100.0%).
• Resurrection: 2 accepted (100.0%).
• Universal sinfulness: 2 accepted (100.0%).
• Salvation through faith: 2 accepted, 1 compromised (83.3%).
• Existence of God: 1 accepted, 1 compromised (75.0%).
• Crucifixion: 3 accepted, 1 refused (75.0%).

Existence of God was mixed. Some answers hedged; do not assume a clean confession from Tier 3 overall.

How this run compares

Against qwen-2.5-7b-instruct (73.0), qwen3.8-omni-flash (72.7), claude-sonnet-5.5 (73.3), gpt-4.1-nano (73.7), perceptron-mk1.5 (73.7), DeepSeek V3.1 Terminus 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 73.0 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: Call to repentance and faith, Evangelistic material, Intercessory prayer.
• Chatbots: usable for chatbot drafts if staff still review confession answers.
• Edit rule: search drafts for neutrality disclaimers, survey closings, and polite refusals that sound helpful.
• Routing: keep DeepSeek V3.1 Terminus 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 17 compromises, patterns show up within a week of real traffic.
• Do not average away Tier scores. Tier 1 at 71.4 can hide a Tier 2 at 71.7 or Tier 3 at 86.7 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.

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.