Reka Edge is a specialized model for visual perception and physical grounding. It is remarkably efficient on image understanding, video analysis, and object detection while remaining par or better in accuracy against other frontier models. This makes it uniquely suited to real-world deployment on highly constrained devices.

With EdgeQ, we compound these advantages with an optimized implementation of Reka Edge for Qualcomm’s Snapdragon 8 Elite, the leading processor in Android phones. This makes it uniquely suited to real-world deployment on highly constrained devices, from phones today to smart glasses, XR headsets, and AI wearables next. It is challenging to retain visual understanding quality and real-time responsiveness within a practical thermal budget, yet EdgeQ delivers state of the art performance in accuracy, speed, and efficiency. We use a Samsung S25 to showcase Qualcomm’s Hexagon NPU below.

Accuracy

EdgeQ (middle column) is the more capable out of box model implementation for your vision tasks. EdgeQ on-device beats even the full-precision server run of Gemma. On RefCOCO and MLVU, the gold-standard benchmarks for object grounding and video understanding, EdgeQ executed on a Samsung S25 phone outperforms Google’s flagship phone model, Gemma 4 E4B (rightmost column), by up to 34 points, on video understanding. EdgeQ uses quantization and processor-specific optimizations to run on-device with small accuracy loss, shown above.

In the table above, Gemma E4B and Reka Edge (leftmost) were run on a server at full precision (FP). This yields a generous best-case accuracy estimate for comparison purposes, even though neither model can run on the S25 with full precision. The speed and efficiency sections use Gemma's only S25-feasible build.

Speed

EdgeQ demonstrates real-time responsiveness on vision understanding, beating Gemma’s latency on both single-image and 6-frame video tasks. We use 6-frame clips since that matches MLVU, the gold standard benchmark, and because it is a feasible temporal resolution to capture real-world behaviors. We were able to achieve more than a 2x speedup by customizing the encoder and decoder separately for the S25’s Qualcomm Hexagon NPU. The times shown above reflect latency from the arrival of the image or clip to the first response of the AI model, in other words, the time to first token (TTFT). This time reflects the model-side delay to first detect presence of an action once a clip is captured.

Although the E4B is Gemma’s flagship Android phone model, it has no on-device NPU path, its only working route on the S25 is the Adreno GPU (native LiteRT-LM). In order to fit on-device, Gemma needs the quantized .litertlm build. The time to first token indicates the time for vision encoding and the prefill to first token, measured on-device, single-tile frames, warm.

Efficiency

Both models received identical 6-frame inputs with preprocessing inside the measured time, from a matched cool start, unplugged, over a fixed 240 s window, display at minimum brightness, with all outputs verified correct.

The practical advantages of EdgeQ widen during continuous perception, the real physical-AI workload. Running our 6-frame inference test from above, back-to-back for 240 s, Gemma’s model brings the peak phone temperature to 71 °C to the touch (skin-sensor zone). Meanwhile, the GPU junction temperature hits a peak of 105 °C, this thermally throttles the processor, degrading turn latency by 36%. The heat build-up slows sustained inference, and by the final 5-inferences in this sustained load benchmark, the average full inference time was 11.6 s per clip. In practice, applications trigger inference on demand or sample a continuous feed at a fixed rate rather than running back to back. Our test deliberately saturates the processor to rigorously expose each stack's ceiling. What matters for a wearable is heat per inference.

On edge devices with limited battery life, EdgeQ can complete about 3x more inferences for the same power consumption over time. A single inference costs about 25 J, a third of the heat of Gemma. For the same test of 240 s of sustained vision, the phone’s case temperature peaked at only 61 °C and the 6-frame latency remained relatively stable with a minimal +15% increase. EdgeQ's implementation keeps the GPU idle, in practice this frees it for what it is specialized for: graphics workloads. This leaves thermal budget decisions to the application in augmented reality settings.

These empirical findings are in part due to a major advantage of the model architecture itself. EdgeQ’s vision encoder, ConvNeXt V2, processes 64 tokens per image tile which allows about 3× more frames per unit of context than comparable VLMs, as explained in our Reka Edge article here. This efficiency in design, paired with kernel engineering for the NPU, quantization, specialized runtimes and a focused parameter count, can sustain real-time video-streaming understanding at the extreme edge.

Built for Snapdragon: the Hexagon NPU Advantage

Our showcase device, the Samsung Galaxy S25, runs Qualcomm's Snapdragon 8 Elite. Its Hexagon NPU delivers roughly 45 TOPS. The Snapdragon Reality Elite, the successor to the Snapdragon XR2 line announced for next-generation smart glasses and XR headsets, carries a 48 TOPS Hexagon NPU from the same architecture family and compute class. EdgeQ's vision and language kernels target the Hexagon NPU directly rather than the GPU. We therefore expect these results to carry over to Reality Elite-class XR devices. The same applies to Snapdragon 8 Elite Gen 5 phones, which ship a newer and faster generation of the same NPU architecture family.

Specialized Deployments

The compute landscape is changing quickly. Extended reality applications are here. EdgeQ is the definitive option, enabling low-latency applications like real-time visual assistance, contextual awareness, and cutting-edge augmented reality on smartphones and smartglasses. Reach out to our team to join a growing list of companies with transformational multimodal deployments that require specialized performance.

Reka Edge, the full precision version of our model, is also available through multiple access options, making it easy to explore and deploy in the environment that best fits your needs.

Get started immediately in the Reka Playground for a hands-on introduction to Reka Edge. For developers seeking to integrate the model into applications, our API documentation provides comprehensive guidance.

For teams requiring self-hosted deployment, Reka Edge can be run locally using your own devices. Please consult the model card for instructions on accessing the model via HuggingFace or deploying it with vLLM.