SAGE (Service Agent) - Property Service: leaderboard

SAGE overall assessment score in the property service customer-service scenario: property service requests such as repairs and noise complaints, combining procedure compliance and conversational quality on a 0-100 scale, higher is better. It is unrelated to the SAGE robustness benchmark.

Metric: Overall Assessment Score (0-100), 0.8 x logical compliance + 0.2 conversational quality, in the Property Service scenario (property service requests such as repairs and noise complaints), over SAGE customer-service dialogues: standard operating procedures formalized as dialogue graphs, an adversarial simulated user with varied intents and personas, turns 1, 5, 10, 15 and the final turn scored; three unnamed judge LLMs (majority vote and mean) plus a deterministic rule engine produce the reference; closed models through their APIs, open models served with vLLM; 0-100 scale; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 27 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Claude Sonnet 4.575.07#267
2Gemini 3 Pro (Preview)73.65#123
3Claude Opus 4.573.25#165
4Kimi K2.571.84#272
5Gemini 3 Flash (Preview)71.62#156
6DeepSeek V3.270.71#425
7DeepSeek R170.58#484
8GLM-4.769.64#383
9Gemini 2.5 Pro69.6#280
10Seed 1.869.03#258
11DeepSeek V367.57#736
12Qwen 3 32B66.71#983
13Qwen 2.5 72B Instruct65.78#1124
14GPT-4.165.46#506
15Qwen 3 14B64.81#1284

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=sage-service-agent-property-service · How It Works · Data refreshed daily, snapshot 2026-11-22.