ESF-Bench - Slot Reset: leaderboard

Metric: Tagged slot accuracy (%; slots tagged with a SR (Slot Reset) scenario, slot and conversation resets and their undoing; 810 synthetic multi-turn enterprise slot-filling samples (6,530 slots, 8 domains) whose prompts combine a schema, a user-assistant conversation and auxiliary sources; varied output formats mapped to a standard form before exact comparison with the ground truth; 32k maximum tokens, reasoning setting as printed per model). Source: arxiv.org. Saturation forecast: Around December 2026. 10 models tracked.

Top models

#ModelScore
1Gemini 2.5 Flash90.1
2GPT-5.1 (High)88.6
3GPT-OSS-120B (High)87
4GPT-5.1 (Non-reasoning)86
5Ministral-3-14B-Reasoning-251284.4
6Qwen 3 32B (Thinking)82.5
7GPT-OSS-120B (Low)80.4
8Gemini 2.5 Flash (Non-reasoning)79.4
9Ministral-3-14B-Instruct-251264.8
10Qwen 3 32B (Non-reasoning)64.4

Interactive version: theaggregate.ai/benchmark?slug=esf-bench-slot-reset · How It Works · Data refreshed daily, snapshot 2026-09-29.