AgentCE-Bench - Meal Planning: leaderboard

AgentCE-Bench (2026) tests tool-using agents on grid planning: the agent fills hidden slots of a weekly schedule by querying static JSON data, with decoy candidates that break global constraints. Task score in percent on the meal planning domain; higher is better.

Metric: Task score (%) on AgentCE-Bench meal domain (54 instances): the agent fills the hidden slots of a 5x7 planning grid with tool calls on static JSON (attribute filter queries, slot and global constraint checkers), each hidden slot having 25 candidates including decoys that violate only the global constraints; every cell is a whole count of instances over the 54 per domain (6 hidden-slot counts by 9 decoy budgets); at most 600 steps; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 13 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 3.5 397B A17B81.5#264
2Qwen 3.5 27B79.6#375
3Qwen 3.5 35B A3B74.1#492
4Qwen 3.5 122B A10B74.1#346
5MiniMax-M2.555.6#600
6MiniMax-M2.153.7#575
7Qwen 3.5 4B46.3#1241
8MiniMax-M246.3#689
9GLM-4.7 FP846.3#335
10Qwen 3.5 9B44.4#828
11AgentCE-Bench MiroThinker-mini 30B (checkpoint unspecified)7.4
12Qwen 3.5 2B1.9#5307
13Qwen 3.5 0.8B0#8232

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

Interactive version: theaggregate.ai/benchmark?slug=agentce-bench-meal-planning · How It Works · Data refreshed daily, snapshot 2026-11-21.