TFLite(LiteRT) - Generative AI on Genio 420

This page lists the generative AI models and representative performance data for Genio 420 platforms. For background information about generative workloads and usage notes, refer to TFLite(LiteRT) - Generative AI.

Model Support and Performance

Note

Performance numbers apply to both Android and Yocto (this platform supports both OS). For OS-level GAI support across all platforms, see AI Supporting Scope.

The following symbols are used in the tables below:

  • -- : To be released.

  • Q3/E : Support planned for Q3 end, 2026.

  • Q4/E : Support planned for Q4 end, 2026.

  • X : Platform does not support this model.

Large Language Models (LLMs)

Performance Comparison (Prompt/Generative) (Unit: tok/s)

Model

Genio 420

Qwen3-0.6B

420.38 / 21.44

Qwen3-1.7B

222.83 / 13.03

Qwen3-4B

105.18 / 6.87

Qwen3-8B

65.04 / 4.58

Qwen2.5-1.5B-Instruct

294.22 / 17.89

Qwen2.5-3B-Instruct

132.30 / 9.98

Qwen2.5-7B-Instruct

57.79 / 4.61

gemma3-1B (Text Only)

513.69 / 24.58

gemma3-4B (Text-Only)

149.00 / 5.69

llama3.2-1B-Instruct

329.32 / 21.58

llama3.2-3B-Instruct

120.17 / 9.92

Phi-3-mini-4k-instruct

101.94 / 6.30

Phi-3.5-mini-instruct

111.01 / 5.30

Vision-Language Models (VLMs)

Performance Comparison (ViT Inference Time / Prompt / Generative)

Model

ViT Inference Time (s)

Prompt Mode (tok/s)

Generative Mode (tok/s)

Qwen3VL-2B

0.49

173.16

13.4803

Speech Recognition Models (ASR)

Performance Data

Model

Genio 420

Whisper-base (8w16a)

Q3/E

Qwen3-ASR-0.6B

–

Moonshine-Tiny

–

Measured latency for these models is reported in asr_cmdline_tool, together with the transcripts each package’s sample clip is expected to produce.

Android-only Models

The following model categories are supported on Android only and are not available on Yocto.

Note

Yocto support for these model categories is planned for 2026/E (estimated).

Stable Diffusion and Image Generation

Performance Comparison (Main Time/Inference Time) (Unit: ms)

Model

Genio 420

Stable Diffusion v2.1 base model with controlnet

39995 / 38425

Stable Diffusion v.1.5 controlnet

43507 / 42111

CLIP and Embedding Models

Performance Comparison (Main Time/Inference Time) (Unit: ms)

Model

Genio 420

img_encoder_proj_clip_vit_large_dynamic

619.67 / 421.55

img_encoder_proj_openclip_vit_h_dynamic

1760.07 / 1392.31

text_encoder_clip_vit_large

408.53 / 58.81

text_encoder_openclip_vit_h

794.72 / 149.17