Encyclo-K - Literature: leaderboard

Encyclo-K's multi-statement selection questions restricted to Literature: accuracy over the 825 Literature questions of the 5,038-question standard set.

Metric: Accuracy on the 825 Literature questions (%). Source: encyclo-k.github.io. Saturation forecast: Around February 2027. 37 models tracked. Updated 17 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Qwen 3 235B A22B 2507 Instruct48.97#618
2Qwen 3 235B A22B 2507 (Thinking)47.76#469
3DeepSeek V3.1 TerminusThinking47.64#440
4Kimi K2 090544.85#564
5Qwen 3 235B A22BThinking43.88#631
6QwQ-32B37.58#908
7Qwen 3 30B A3B 2507 (Thinking)37.45#765
8Qwen 3 32BThinking35.76#965
9Qwen 3 30B A3B 2507 Instruct35.03#1091
10Qwen 3 30B A3BThinking32.48#1155
11DeepSeek V3 (0324)32.24#722
12Qwen 3 14BReasoning28.73#1264
13Qwen 2.5 32B Instruct27.64#1257
14Qwen 2.5 72B Instruct26.55#1119
15Qwen 3 4B 2507 Instruct26.3#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-literature · How It Works · Data refreshed daily, snapshot 2026-11-17.