# The Multivac — Evaluation Report

**Evaluation ID:** EVAL-20260207-154722
**Date:** Feb 21, 2026
**Category:** edge cases
**Question ID:** EDGE-006

---

## Question

Calculate and explain any issues with:

1. 0.1 + 0.2 = ?
2. 2^53 + 1 in JavaScript
3. 1/3 represented as a finite decimal
4. sqrt(-1) in Python without importing cmath
5. 10^309 in most programming languages
6. What's the result of: (-1) % 10 in Python vs JavaScript?

Don't just give answers - explain WHY these are problematic.

---

## Winner

**Claude Sonnet 4.5** (Anthropic)
- Winner Score: 9.83
- Matrix Average: 9.35
- Total Judgments: 90

---

## Rankings

| Rank | Model | Provider | Avg Score | Judgments |
|------|-------|----------|-----------|----------|
| 1 | Claude Sonnet 4.5 | Anthropic | 9.83 | 9 |
| 2 | Claude Opus 4.5 | Anthropic | 9.81 | 8 |
| 3 | Grok 4.1 Fast | xAI | 9.78 | 7 |
| 4 | GPT-5.2-Codex | OpenAI | 9.55 | 8 |
| 5 | Grok 3 (Direct) | xAI | 9.54 | 7 |
| 6 | DeepSeek V3.2 | DeepSeek | 9.49 | 8 |
| 7 | Gemini 3 Flash Preview | Google | 9.43 | 8 |
| 8 | MiMo-V2-Flash | Xiaomi | 9.41 | 8 |
| 9 | GPT-OSS-120B | OpenAI | 8.99 | 8 |
| 10 | Gemini 3 Pro Preview | Google | 7.67 | 8 |

---

## 10×10 Judgment Matrix

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

| Judge ↓ / Resp → | Claude Opus | Gemini 3 | Claude Sonnet | GPT-5.2-Codex | GPT-OSS-120B | Gemini 3 | DeepSeek V3.2 | MiMo-V2-Flash | Grok 4.1 Fast | Grok 3 |
|---|---|---|---|---|---|---|---|---|---|---|
| Claude Opus | — | 7.5 | 9.8 | 9.3 | 8.8 | 9.8 | 9.4 | 9.6 | 10.0 | 9.8 |
| Gemini 3 | 10.0 | — | 10.0 | 0.0 | 0.0 | 0.0 | 9.8 | 0.0 | 0.0 | 0.0 |
| Claude Sonnet | 10.0 | 8.3 | — | 9.8 | 9.8 | 9.8 | 9.0 | 10.0 | 9.8 | 9.8 |
| GPT-5.2-Codex | 9.0 | 5.0 | 9.8 | — | 7.5 | 8.8 | 9.4 | 8.0 | 9.2 | 8.6 |
| GPT-OSS-120B | 0.0 | 0.0 | 9.4 | 9.3 | — | 9.2 | 0.0 | 8.8 | 0.0 | 0.0 |
| Gemini 3 | 10.0 | 8.1 | 10.0 | 9.8 | 9.6 | — | 9.8 | 9.6 | 10.0 | 9.8 |
| DeepSeek V3.2 | 10.0 | 8.8 | 10.0 | 9.1 | 9.2 | 9.8 | — | 9.8 | 9.8 | 9.8 |
| MiMo-V2-Flash | 9.8 | 8.6 | 9.8 | 9.8 | 9.0 | 8.6 | 9.8 | — | 10.0 | 9.0 |
| Grok 4.1 Fast | 10.0 | 7.5 | 10.0 | 10.0 | 9.2 | 10.0 | 9.8 | 10.0 | — | 10.0 |
| Grok 3 | 9.7 | 7.7 | 9.6 | 9.4 | 8.8 | 9.4 | 8.8 | 9.7 | 9.7 | — |

---

## 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: EDGE-006. 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-20260207-154722/results
Full dataset: https://app.themultivac.com/dashboard/export