MuSeToM - Task 4 - F1: leaderboard
MuSeToM - Task 4 - F1 — 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: Macro-F1 (%).
Metric: Macro-F1 (%). Source: arxiv.org. Saturation forecast: Around June 2028. 18 models tracked. Updated 15 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 397B A17B | 71.24 | #251 |
| 2 | MiniMax-M2.1 | 71.07 | #574 |
| 3 | GPT-OSS-120B | 70.88 | #703 |
| 4 | MiniMax-M2.5 | 70.74 | #584 |
| 5 | Kimi K2 (Thinking) | 69.45 | |
| 6 | QwQ-32B | 67.33 | #915 |
| 7 | DeepSeek R1 | 67.1 | #478 |
| 8 | Qwen 3 235B A22B | 66.73 | #628 |
| 9 | Phi-3-medium-4k-instruct | 66.37 | #2696 |
| 10 | Llama 4 Maverick Instruct | 64.19 | #1163 |
| 11 | GPT-3.5 Turbo | 62.79 | #2973 |
| 12 | DeepSeek V3.2 | 60.48 | #412 |
| 13 | GPT-4o | 59.54 | #744 |
| 14 | Llama 3.1 405B Instruct | 59.15 | #1122 |
| 15 | Mistral 7B Instruct (v0.3) | 58.61 | #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-f1 · How It Works · Data refreshed daily, snapshot 2026-11-16.