Nejumi 4 - HalluLens: leaderboard
HalluLens hallucination test as run in Nejumi Leaderboard 4 (Weights & Biases Japan): prompts about nonexistent entities, with GPT-4.1 judging whether the model refuses or hallucinates. Score is Nejumi's hallucination_resistance rate (%); higher is better.
Metric: Hallucination resistance (%). Source: nejumi.ai. Saturation forecast: Not forecast. 145 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 (Thinking) | 100 | #148 (Claude Sonnet 4.5) |
| 2 | O3 (2025-04-16) | 100 | #125 |
| 3 | Claude Opus 5 (Max) | 100 | #16 (Claude Opus 5) |
| 4 | Claude Opus 4.5 (Thinking) | 100 | #87 (Claude Opus 4.5) |
| 5 | Claude Sonnet 4.6 (Thinking) | 100 | #94 (Claude Sonnet 4.6) |
| 6 | Claude Opus 4.5 (Non-reasoning) | 100 | #87 (Claude Opus 4.5) |
| 7 | Claude Opus 4.7 (xHigh) | 100 | #51 (Claude Opus 4.7) |
| 8 | Claude Opus 4.8 (xHigh) | 100 | #39 (Claude Opus 4.8) |
| 9 | Claude Sonnet 4 (Non-reasoning) | 100 | #207 (Claude Sonnet 4) |
| 10 | Kimi K3 (Thinking) | 100 | #35 (Kimi K3) |
| 11 | Claude Opus 4 (Non-reasoning) | 100 | #170 (Claude Opus 4) |
| 12 | Kimi K3 (Max) | 99 | #35 (Kimi K3) |
| 13 | Claude Opus 4.6 (Thinking) | 99 | #67 (Claude Opus 4.6) |
| 14 | Claude Haiku 4.5 (Thinking) | 99 | #287 (Claude Haiku 4.5) |
| 15 | Solar Pro 4 | 99 | #230 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=nejumi-4-hallulens · How It Works · Data refreshed daily, snapshot 2026-11-05.