X-RIC BARX - Schema Compliance: leaderboard
BARX v1 TableI source-local O-RAN TLM candidates: L3 LLaMA3-8B; Q7B Qwen7B; Q4B Qwen4B; QT Qwen:latest; MT Mistral:latest; TL TinyLLaMA:latest; DS DeepSeek7B. SectionII frozenLLM+trained256Dstateencoder versus SectionIV fine-tuned wording unresolved; retain paper-specific trained system identities and never map to stock models. Three regimes(T,p)=(0.01,0.8),(0.1,0.9),(0,1), five20query trials each,300responses/model. Latency and pairwise determinism not capability scores. Metric: Draft-7 JSON-schema compliance (%). Higher is better. Source: https://arxiv.org/html/2610.22269v1
Metric: Draft-7 JSON-schema compliance (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | X-RIC TLM L3 (BARX paper2610.22269) | 100 | |
| 2 | X-RIC TLM Q7B (BARX paper2610.22269) | 100 | |
| 3 | X-RIC TLM Q4B (BARX paper2610.22269) | 100 | |
| 4 | X-RIC TLM MT (BARX paper2610.22269) | 100 | |
| 5 | X-RIC TLM DS (BARX paper2610.22269) | 100 | |
| 6 | X-RIC TLM QT (BARX paper2610.22269) | 83.3 | |
| 7 | X-RIC TLM TL (BARX paper2610.22269) | 0 |
Interactive version: theaggregate.ai/benchmark?slug=x-ric-barx-schema-compliance · How It Works · Data refreshed daily, snapshot 2026-10-17.