mindfossil/telecom-intelligence-model-v7-merged

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7.6BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jul 31, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The mindfossil/telecom-intelligence-model-v7-merged is a 7.6 billion parameter language model built on Qwen2.5-7B-Instruct, fine-tuned for telecom network operations. It excels at root cause analysis, KPI degradation diagnosis, and generating CLI commands across multi-vendor RAN/Core nodes with a 32768 token context length. This model is specifically optimized for deep knowledge in 5G NR, LTE RAN, and 5G Core NF attribution, making it highly specialized for telecommunications intelligence tasks.

Loading preview...

Telecom Intelligence Model v7 Overview

This model, mindfossil/telecom-intelligence-model-v7-merged, is a 7.6 billion parameter language model based on Qwen2.5-7B-Instruct, specifically fine-tuned for telecom network operations. It was trained using QLoRA (4-bit) and SFT via Unsloth, incorporating 805 training examples, with 200 new examples added in v7 covering 5G NR internals, multi-vendor PM counter diagnosis, and advanced 5GC NF fault chains. The model is provided as a merged, standalone artifact, compatible with vLLM, Transformers, and HuggingFace Inference Endpoints.

Key Capabilities

  • Root Cause Analysis: Diagnoses KPI degradations from PM counter data across Ericsson, Huawei, Nokia, and ZTE RAN/Core nodes.
  • Deep Telecom Knowledge: Possesses extensive knowledge of 5G NR (e.g., SSB beam management, numerology, CORESET/PDCCH), LTE RAN (KPI chains, HO parameters), and 5G Core NF attribution (identifying specific NFs and interfaces).
  • Multi-vendor Support: Handles counter normalization for Ericsson, Huawei, Nokia, and ZTE, and generates canonical Huawei MML and Ericsson AMOS CLI commands.
  • Operational Data Reasoning: Reasons step-by-step over telecom operational data to identify anomalies, diagnose faults, and recommend actions.

Good For

  • Network Engineers: Augmenting troubleshooting and diagnosis workflows for 4G/5G networks.
  • Automated Operations: Integrating into systems for automated anomaly detection and preliminary root cause identification.
  • Training & Development: Understanding complex telecom concepts and generating relevant CLI commands for various vendors.

Finetunes
1 models