BudHist - 2026-08 OpenRouter: leaderboard
Four-option Buddhist-history MCQ; 2026-08 edition with250questions. Native numerator retained instead of rounded printed percentage. Provider/deployment contexts separate; no internet-search claim inferred. Blank/X entries unmeasured and not zeros. Local quantization retains literal Q4_K_M; local historical tags lacking exact checkpoint lineage withheld.
Metric: Correct answers. Source: mbingenheimer.net. Saturation forecast: Estimated already saturated. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5.2 | 238 | #80 |
| 2 | MiniMax-M3 | 234 | #151 |
| 3 | Claude Sonnet 4.6 | 233 | #94 |
| 4 | Gemini 3.6 Flash | 233 | #57 |
| 5 | DeepSeek V4 Pro | 232 | #112 |
| 6 | Kimi K2 (Thinking) | 231 | #252 (Kimi K2) |
| 7 | Inkling | 231 | #144 |
| 8 | Claude Sonnet 4.5 | 230 | #148 |
| 9 | Kimi K2.6 | 230 | #113 |
| 10 | DeepSeek V4 Flash | 230 | #155 |
| 11 | Gemini 3.5 Flash Lite | 229 | #174 |
| 12 | GPT-5.6 Luna | 222 | #79 |
| 13 | Inkling Small | 221 | #188 |
| 14 | Qwen 3.6 35B A3B | 217 | #238 |
| 15 | GPT-4.1 | 216 | #258 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=budhist-2026-08-openrouter · How It Works · Data refreshed daily, snapshot 2026-11-05.