TeleCom-Bench - Core Network: leaderboard
Product-knowledge set of TeleCom-Bench (Xiao et al., ZTE Corporation, KDD 2026): open-ended questions on core-network functions and platforms (PCF, UDM, AMF, NFV orchestration, distributed storage and charging gateways), built from vendor documentation with a knowledge graph. Three LLM judges vote on a 5-point scale, printed as a percentage (one set), with reasoning enabled, temperature 0.7 and a majority vote over three samples; higher is better.
Metric: LLM-judge score (%; open-ended core-network product-knowledge questions, three-judge vote on a 5-point scale). Source: arxiv.org. Saturation forecast: Around February 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V3.2 (Thinking) | 62.34 | #403 (DeepSeek V3.2) |
| 2 | Qwen 3 235B A22B (Thinking) | 62.13 | #622 (Qwen 3 235B A22B) |
| 3 | Qwen 3 32B (Thinking) | 60.22 | #960 (Qwen 3 32B) |
| 4 | GLM-4.7 (Thinking) | 47.68 | #365 (GLM-4.7) |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=telecom-bench-core-network · How It Works · Data refreshed daily, snapshot 2026-11-12.