Skip to content

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

docs(model_selection): lead with flagship Fun-ASR-Nano (#3020)

zhifu gao2ad0514430c8
diff --git a/docs/model_selection.md b/docs/model_selection.md
index 7dfc6daba..cd1508ef8 100644
--- a/docs/model_selection.md
+++ b/docs/model_selection.md
@@ -4,7 +4,17 @@ Use this guide when you are choosing a first model, comparing FunASR with Whispe
## Fast default path
-If you are unsure, start with **SenseVoice-Small**:
+If you have a GPU, start with the flagship **Fun-ASR-Nano** — an LLM-based ASR model (SenseVoice encoder + a Qwen3 decoder) covering 31 languages, with the strongest accuracy on hard cases, context, and proper nouns:
+
+```python
+from funasr import AutoModel
+
+model = AutoModel(model="FunAudioLLM/Fun-ASR-Nano-2512", device="cuda")
+result = model.generate(input="meeting.wav")
+print(result[0]["text"])
+```
+
+On CPU, or when you want multilingual + emotion/event tags and speaker-aware meeting transcripts in one fast non-autoregressive pass, use **SenseVoice-Small**:
```python
from funasr import AutoModel
@@ -18,7 +28,7 @@ model = AutoModel(
result = model.generate(input="meeting.wav")
```
-It is the best first choice for demos, private APIs, multilingual transcription, speaker-aware meeting transcripts, and agent voice input. Switch only when your workload has a clear requirement such as Mandarin production accuracy, streaming latency, or LLM-based ASR experiments.
+Switch to Paraformer when your workload is Mandarin-only and you want character-level timestamps or hotwords.
## Decision table
@@ -79,9 +89,9 @@ For migration work, use the [migration benchmark example](../examples/migration/
## Practical recommendations
-- Start with SenseVoice-Small for demos, private APIs, agent voice input, and multilingual workloads.
-- Use Paraformer when your production traffic is primarily Mandarin and you want the mature non-autoregressive ASR path.
-- Use Fun-ASR-Nano when you specifically want the LLM-based model path or vLLM acceleration experiments.
+- With a GPU, default to Fun-ASR-Nano — the flagship LLM-based model (31 languages), strongest on hard, contextual, and proper-noun-heavy audio.
+- On CPU, or for multilingual + emotion workloads, use SenseVoice-Small (fast non-autoregressive, CPU-viable).
+- Use Paraformer when your production traffic is primarily Mandarin and you want timestamps or hotwords.
- Use the streaming runtime when partial results and long-lived connections matter more than a single final transcript.
- Keep model aliases stable in production runbooks so benchmark results and bug reports are reproducible.
- Open a [Deployment Help issue](https://github.com/modelscope/FunASR/issues/new?template=deployment_help.md) with model, device, command, logs, audio duration, and runtime path when you get stuck.
diff --git a/docs/model_selection_zh.md b/docs/model_selection_zh.md
index dc883e560..ae5208e06 100644
--- a/docs/model_selection_zh.md
+++ b/docs/model_selection_zh.md
@@ -4,7 +4,17 @@
## 默认快速路径
-如果还不确定,先从 **SenseVoice-Small** 开始:
+如果有 GPU,先从旗舰 **Fun-ASR-Nano** 开始 —— 基于 LLM 的识别模型(SenseVoice 编码器 + Qwen3 解码),覆盖 31 语种,在难例、上下文和专名上精度最强:
+
+```python
+from funasr import AutoModel
+
+model = AutoModel(model="FunAudioLLM/Fun-ASR-Nano-2512", device="cuda")
+result = model.generate(input="meeting.wav")
+print(result[0]["text"])
+```
+
+在 CPU 上,或当你想要多语种 + 情感/事件标签、带说话人信息的会议转写一次完成时,用 **SenseVoice-Small**:
```python
from funasr import AutoModel
@@ -18,7 +28,7 @@ model = AutoModel(
result = model.generate(input="meeting.wav")
```
-它适合 demo、私有 API、多语种转写、带说话人信息的会议转写和 Agent 语音输入。只有当你的场景明确需要中文生产识别、低延迟流式结果或 LLM-based ASR 实验时,再切换到其他路径。
+当你的场景是纯中文、需要字级时间戳或热词时,切换到 Paraformer。
## 决策表