Encyclo-K - History: leaderboard

Encyclo-K's multi-statement selection questions restricted to History: accuracy over the 686 History questions of the 5,038-question standard set.

Metric: Accuracy on the 686 History questions (%). Source: encyclo-k.github.io. Saturation forecast: Around October 2027. 37 models tracked. Updated 17 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Qwen 3 235B A22B 2507 (Thinking)45.19#469
2DeepSeek V3.1 TerminusThinking45.04#440
3Qwen 3 235B A22B 2507 Instruct43.59#618
4Kimi K2 090540.96#564
5Qwen 3 235B A22BThinking37.9#631
6Qwen 3 30B A3B 2507 (Thinking)34.99#765
7QwQ-32B34.84#908
8Qwen 3 32BThinking32.65#965
9Qwen 3 30B A3B 2507 Instruct32.51#1091
10DeepSeek V3 (0324)29.88#722
11Qwen 3 30B A3BThinking28.43#1155
12Qwen 3 14BReasoning26.68#1264
13Qwen 2.5 32B Instruct25.22#1257
14Qwen 3 8BThinking25.22#1868
15Qwen 3 4B 2507 Instruct24.93#2215

No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.

Interactive version: theaggregate.ai/benchmark?slug=encyclo-k-history · How It Works · Data refreshed daily, snapshot 2026-11-17.