RF-GPT: Teaching AI to See the Wireless World

Hang Zou; Yu Tian; Bohao Wang; Lina Bariah; Samson Lasaulce; Chongwen Huang; Mérouane Debbah

eess.SPFeb 16, 2026

RF-GPT: Teaching AI to See the Wireless World

Authors: Hang Zou, Yu Tian, Bohao Wang, Lina Bariah, Samson Lasaulce, Chongwen Huang, Mérouane Debbah

Organizations: Research Institute for Digital Future, Khalifa University, 127788 Abu Dhabi, UAE · College of Information Science and Electronic Engineering, Zhejiang University, 310027, Hangzhou, China · Université de Lorraine, CNRS, CRAN, F-54000 Nancy, France

Abstract

Large language models (LLMs) and multimodal models have become powerful general-purpose reasoning systems. However, radio-frequency (RF) signals, which underpin wireless systems, are still not natively supported by these models. Existing LLM-based approaches for telecom focus mainly on text and structured data, while conventional RF deep-learning models are built separately for specific signal-processing tasks, highlighting a clear gap between RF perception and high-level reasoning. To bridge this gap, we introduce RF-GPT, a radio-frequency language model (RFLM) that utilizes the visual encoders of multimodal LLMs to process and understand RF spectrograms. In this framework, complex in-phase/quadrature (IQ) waveforms are mapped to time-frequency spectrograms and then passed to pretrained visual encoders. The resulting representations are injected as RF tokens into a decoder-only LLM, which generates RF-grounded answers, explanations, and structured outputs. To train RF-GPT, we perform supervised instruction fine-tuning of a pretrained multimodal LLM using a fully synthetic RF corpus. Standards-compliant waveform generators produce wideband scenes for six wireless technologies, from which we derive time-frequency spectrograms, exact configuration metadata, and dense captions. A text-only LLM then converts these captions into RF-grounded instruction-answer pairs, yielding roughly 120,000 RF scenes and 0.625 million instruction examples without any manual labeling. Across benchmarks for wideband modulation classification, overlap analysis, wireless-technology recognition, WLAN user counting, and 5G NR information extraction, RF-GPT achieves strong multi-task performance, whereas general-purpose VLMs with no RF grounding largely fail.

Figures & tables

Model / family	Input modality	Backbone	Training paradigm	Heads?	Output format
WFM [ 6 ]	IQ / channel	Transformer encoder	SSL + task heads	yes	per-head logits
LWM [ 7 ]	CSI	Transformer encoder	SSL + task heads	yes	per-head logits
WirelessGPT [ 8 ]	CSI	Transformer encoder	Pretrain + FT	yes	per-head logits
WiFo [ 9 ]	CSI	Transformer encoder	SSL (MAE)	no	numeric outputs
LLM4WM [ 13 ]	CSI	LLM backbone	Multi-task SFT	yes	numeric outputs
WirelessAgent [ 14 ]	KPIs / network	LLM backbone	RAG / tool-use	no	text + actions

TABLE I: Positioning of RF-GPT against related works.

Fig. 2: Wireless technologies gallery: time-frequency spectrograms from six important RF signals including 5G NR, 4G LTE, 3G UMTS, WLAN, DVB-S2 and Bluetooth.

Fig. 3: Our fine-grained caption strategy for RF spectrograms.

Fig. 4: RF-grounded instruction synthesis pipeline.

Benchmark	Random	Majority
WTR	5.27	12.70
NRIE	23.66	29.00
WBOD	19.44	41.33
WBMC	1.11	3.30
WNUC	9.60	27.32

TABLE II: Random-guess and majority-class baseline accuracies (%), averaged across difficulty splits.

Fig. 5: Comparison between Qwen2.5-VL-7B-Instruct and RF-GPT on a 5G DL spectrogram with only one UE.

Fig. 6: Benchmark scores on the wide‑band modulation classification (WBMC) tasks.

Fig. 7: Benchmark scores on the wide‑band overlap detection (WBOD) tasks.

Fig. 8: Benchmark results for wireless technology recognition (WTR) benchmark.

Fig. 9: Benchmark scores on the WLAN number of users counting (WNUC) tasks for 802.11ax and 802.11be.

Fig. 10: Benchmark results for the New Radio information extraction (NRIE) benchmark.

Failure mode	Easy	Medium	Hard
Fully correct	53.80	40.60	9.20
Count error (under-predict)	2.00	2.05	2.15
Count error (over-predict)	0.30	0.20	0.50
Wrong temporal ordering	0.80	4.90	0.30
Per-signal class confusion	43.10	52.25	87.85

TABLE III: WBMC error decomposition on RF-GPT-7B.

Fig. 11: WTR confusion matrix on RF-GPT-7B.

Component	RF-GPT-3B	RF-GPT-7B
STFT (CPU)	0.65 ms	0.95 ms
Image preproc. (CPU)	51.34 ms	51.78 ms
Vision encoder (GPU)	31.64 ms	32.03 ms
LLM prefill (GPU, derived)	16.00 ms	18.05 ms
LLM decode (GPU, per token)	3.07 ms	4.79 ms
End-to-end (1 img + 64 tok)	292.93 ms	404.71 ms

TABLE IV: Inference latency and peak HBM on NVIDIA H200.

Method	Easy	Medium	Hard
Oracle-Loc + ResNet18	99.43	63.61	63.61
YOLOv8s det+cls	89.49	24.77	24.77
RF-GPT-3B	80.00	71.40	43.69
RF-GPT-7B	82.41	74.24	47.77

TABLE V: WBMC baselines vs RF-GPT.

Fig. 12: NRIE ablation vs. impairment level of IQ imbalance (IQ), power amplifier nonlinearity (PA), tapped delay line (TDL), and carrier frequency offset (CFO).

Fig. 13: NRIE ablation comparing RF-GPT and Transformer/CNN baselines across epochs.

Configuration	WTR	NRIE	WBMC	WBOD	WNUC
LLM-only	99.64	76.96	68.14	83.59	70.17
Vision-only	99.79	72.84	73.76	84.66	68.18
Full-FT	99.88	74.13	75.15	89.73	72.47
Grayscale	89.49	68.73	51.63	79.43	49.45
25 % data	99.28	64.63	57.42	71.27	55.03
50 % data	99.71	68.93	62.85	77.69	63.60

TABLE VI: Design-choice ablations on RF-GPT-7B.

Fig. 14: WBMC accuracy of RF-GPT-7B under different image resolutions.

Metric	RF-GPT	Qwen2.5-VL	Llama 3.2-V
PAES, in-dist. OTA	96	32	60
PAES, OOD low-SNR (shielded)	40	0	0
Prompt-leakage rate	15	55	70

TABLE VII: Sim-to-real transfer on live SDR captures [ 36 ] .

UTB question	RF-GPT-7B	RF-GPT-3B	Qwen2.5-VL-7B
Q1 Coexistence detection	62.8	49.6	48.7
Q2 Mod.-diversity count (within one)	98.8	98.8	56.7
Q3 Burst pattern (scene)	100.0	58.0	50.0

TABLE VIII: Unseen-Task Benchmark accuracy (%).

WBMC split	Magnitude	Magnitude-Phase (Coherent)	Magnitude-Phase (Polar)
WBMC-Easy	63.04	26.38 ( $-36.66$ )	52.49 ( $-10.55$ )
WBMC-Medium	47.43	13.82 ( $-33.61$ )	36.21 ( $-11.22$ )
WBMC-Hard	22.54	5.52 ( $-17.02$ )	16.26 ( $-6.28$ )

TABLE IX: Magnitude vs phase-aware encodings, WBMC signal accuracy (%);

\Delta

in parentheses is vs Magnitude. RF-GPT-3B fine-tuned for 1 epoch under identical hyperparameters; only the per-pixel encoding differs across the three configurations.

Wireless Communications

Wireless communications studies how information is sent over radio links and how networks manage signals, spectrum, and resources reliably.

See all 391 papers

Cross-Modal Generation: From Commodity WiFi to High-Fidelity mmWave and RFID Sensing
Apr 17, 2026Zhixiong Yang, Long Jing, Yao Li +3Synthetic Data Generation Wireless Communications
Tsinghua University Beijing, China
RF-Agent: A Practical Framework for Building Language Agents for RFIC Design
Jul 21, 2026Yueqi Xing, Houbo He, Jolie Wang +5Retrieval-Augmented Generation LLM Fine-Tuning
ECE Department, Rice University, Houston, TX, USA · ECE Department, The George Washington University, Washington, D.C., USA
TelecomGPT-R1: Unified Post-Training for Reasoning Across Heterogeneous Telecom Tasks
Sep 21, 2026Bohao Wang, Chenwei Wu, Hang Zou +7LLM Grounding RL for Language Model Reasoning
College of Information Science and Electronic Engineering, Zhejiang University, 310027 Hangzhou, China · Research Institute for Digital Future, Khalifa University, P O Box 127788, Abu Dhabi, UAE · Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, MI 48109-2122, USA +1

Apr 17, 2026J/K move · Enter open · S save

Cross-Modal Generation: From Commodity WiFi to High-Fidelity mmWave and RFID Sensing

Zhixiong Yang, Long Jing, Yao Li, Shuli Cheng+2

Tsinghua University Beijing, China

Synthetic Data Generation Wireless Communications Multimodal Generation Multimodal Diffusion Models

Open paper AskPDF arXiv

AIGC has shown remarkable success in CV and NLP, and has recently demonstrated promising potential in the wireless domain. However, significant data imbalance exists across RF modalities, with abundant WiFi data but scarce mmWave and RFID data due to high acquisition cost. This makes it difficult to train high-quality generative models for these data-scarce modalities. In this work, we propose RF-CMG, a diffusion-based cross-modal generative method that leverages data-rich WiFi signals to synthesize high-fidelity RF data for scarce modalities including mmWave and RFID. The key insight of RF-CMG is to decouple cross-modal generation into high-frequency guidance and low-frequency constraint, which respectively learn high-frequency distribution from limited target modality data and preserve the underlying physical structure via low-frequency constraints during generation. On this basis, we introduce a Modality-Guided Embedding (MGE) module to steer the reverse diffusion trajectory toward the target high-frequency distribution, and a Low-Frequency Modality Consistency (LFMC) module to progressively enforce low-frequency constraints to suppress the accumulation of source-modality structural biases during inference, enabling high-quality target-modality generation. Performance comparison with several prevalent generative models demonstrates that RF-CMG achieves superior performance in synthesizing RFID and mmWave signals. We further showcase the effectiveness of the data generated by RF-CMG in gesture recognition tasks, and analyze the impact of the proportion of synthetic data on downstream performance.

Abstract

Wireless Communications

Explore similar work

Cross-Modal Generation: From Commodity WiFi to High-Fidelity mmWave and RFID Sensing

RF-Agent: A Practical Framework for Building Language Agents for RFIC Design

TelecomGPT-R1: Unified Post-Training for Reasoning Across Heterogeneous Telecom Tasks