# The Multivac — Evaluation Report

**Evaluation ID:** EVAL-20260402-231136
**Date:** Apr 02, 2026
**Category:** communication
**Question ID:** COMM-016

---

## Question

Explain how HTTPS works to someone who only knows that 'the lock icon means secure.' Cover: (1) What actually happens when you type https://example.com, (2) Why you need certificates, (3) What a man-in-the-middle attack is, (4) Why public WiFi is risky even with HTTPS. Use analogies only — no technical terms (no 'encryption,' 'certificate,' 'handshake'). Max 400 words.

---

## Winner

**Claude Opus 4.6** (openrouter)
- Winner Score: 9.00
- Matrix Average: 8.11
- Total Judgments: 87

---

## Rankings

| Rank | Model | Provider | Avg Score | Judgments |
|------|-------|----------|-----------|----------|
| 1 | Claude Opus 4.6 | openrouter | 9.00 | 8 |
| 2 | GPT-5.4 | openrouter | 8.75 | 9 |
| 3 | Claude Sonnet 4.6 | openrouter | 8.70 | 8 |
| 4 | Grok 4.20 | openrouter | 8.69 | 8 |
| 5 | GPT-OSS-120B | OpenAI | 8.38 | 9 |
| 6 | DeepSeek V4 | openrouter | 8.08 | 9 |
| 7 | MiMo-V2-Flash | Xiaomi | 8.02 | 9 |
| 8 | Gemini 3.1 Pro | openrouter | 7.41 | 9 |
| 9 | Mistral Small Creative | Mistral | 7.31 | 9 |
| 10 | Seed 1.6 Flash | openrouter | 6.80 | 9 |

---

## 10×10 Judgment Matrix

Rows = Judge, Columns = Respondent. Self-judgments excluded (—).

| Judge ↓ / Resp → | GPT-OSS-120B | Claude Opus | GPT-5.4 | Claude Sonnet | Gemini 3.1 Pro | Grok 4.20 | DeepSeek V4 | MiMo-V2-Flash | Mistral Small | Seed 1.6 Flash |
|---|---|---|---|---|---|---|---|---|---|---|
| GPT-OSS-120B | — | 8.7 | 8.4 | 8.2 | 7.8 | 8.7 | 8.0 | 8.7 | 7.1 | 8.3 |
| Claude Opus | 7.3 | — | 8.8 | 8.6 | 7.5 | 8.3 | 6.5 | 7.3 | 7.0 | 5.5 |
| GPT-5.4 | 7.8 | 9.0 | — | 8.0 | 5.8 | 9.0 | 8.2 | 8.4 | 6.8 | 5.8 |
| Claude Sonnet | 8.2 | 9.0 | 8.0 | — | 7.0 | 8.3 | 7.5 | 7.8 | 6.4 | 5.0 |
| Gemini 3.1 Pro | 8.1 | 9.8 | 9.8 | · | — | 8.2 | 8.2 | 6.5 | 4.8 | 4.2 |
| Grok 4.20 | 8.3 | 8.3 | 8.3 | 9.0 | 6.3 | — | 8.3 | 7.4 | 7.5 | 6.3 |
| DeepSeek V4 | 9.0 | · | 8.6 | 9.0 | 8.8 | 8.6 | — | 8.6 | 9.0 | 8.8 |
| MiMo-V2-Flash | 9.0 | 9.2 | 9.0 | 9.2 | 8.6 | 9.0 | 9.0 | — | 9.0 | 8.6 |
| Mistral Small | 9.6 | 9.4 | 9.4 | 9.4 | 8.8 | 9.4 | 9.0 | 9.4 | — | 8.8 |
| Seed 1.6 Flash | 8.2 | 8.7 | 8.4 | 8.2 | 6.0 | · | 8.0 | 8.0 | 8.2 | — |

---

## Methodology

- **10×10 Blind Peer Matrix:** All models answer the same question, then all models judge all responses.
- **5 Criteria:** Correctness, completeness, clarity, depth, usefulness (each scored 1–10).
- **Self-judgments excluded:** Models do not judge their own responses.
- **Weighted Score:** Composite of all 5 criteria.

---

## Citation

The Multivac (2026). Blind Peer Evaluation: COMM-016. app.themultivac.com

## License

Open data. Free to use, share, and build upon. Please cite The Multivac when using this data.

Download raw JSON: https://app.themultivac.com/api/evaluations/EVAL-20260402-231136/results
Full dataset: https://app.themultivac.com/dashboard/export