MuSeToM - Task 4 - Accuracy: leaderboard

MuSeToM - Task 4 - Accuracy — MuSeToM v1 Table1 direct prompting on human-validated textual belief revision; clean tasks1binary/2multiplechoice and noisy tasks3binary/4multiplechoice remain separate. Accuracy and Macro-F1 are distinct percentage metrics. AppendixB.1 binds the Llama/Mistral/Mixtral instruction checkpoints; Tables23and28–31bind the sole Phi row to Phi-3-Medium-4K-Instruct in the same18-model cohort. temperature0,top_p1,1024output tokens for the listed first group and10240for later reasoning models; Phi output budget is unreported because B.1omits it. Effort/API dated snapshots are unreported. Eighteen model rows; Human reference is a separate reference, not Median Human/Expert and not an AI model. Main table shorthand is resolved from this paper only; no global Base/Instruct or generation alias. Metric: Accuracy (%).

Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around June 2028. 18 models tracked. Updated 15 Nov 2026.

Top models

#ModelScoreOverall rank
1Qwen 3.5 397B A17B71.27#251
2MiniMax-M2.171.07#574
3GPT-OSS-120B70.94#703
4MiniMax-M2.570.73#584
5Kimi K2 (Thinking)69.45
6QwQ-32B67.16#915
7DeepSeek R167.15#478
8Qwen 3 235B A22B66.76#628
9Phi-3-medium-4k-instruct66.42#2696
10Llama 4 Maverick Instruct64.33#1163
11GPT-3.5 Turbo62.91#2973
12DeepSeek V3.260.62#412
13GPT-4o60.01#744
14Llama 3.1 405B Instruct59.27#1122
15Mistral 7B Instruct (v0.3)58.53#6719

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=musetom-task-4-accuracy · How It Works · Data refreshed daily, snapshot 2026-11-16.