# Ming-3.0-tts-tiny Benchmark Scores & Performance

> Ming-3.0-tts-tiny has inclusionAI Ming-3.0 results results. Voice evidence is display-only and does not produce a text-model score.

Canonical page: https://benchlm.ai/models/ming-3-0-tts-tiny

Last updated: October 20, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | InclusionAI |
| Source Type | Open Weight |
| Reasoning Type | Non-Reasoning |
| Context Window | N/A |
| Official model card | [InclusionAI model documentation](https://huggingface.co/inclusionAI/Ming-3.0-tts-tiny) |
| Overall Score | Not scored (voice evidence only) |
| Overall Rank | Unranked |

## Family & Coverage

- Family: Ming-3.0-tts-tiny
- Variant: base
- Benchmarks covered: 0 of 670
- Coverage note: Speech and voice results appear below; no weighted text-model benchmark rows are stored.

## Speech and voice evidence

[Ming-3.0 speech evaluation](/voice-benchmarks/ming-3-0-speech-evaluation)

All results are inclusionAI-run or developer-reported by inclusionAI. Text-to-speech figures use a unified ASR-based WER or CER protocol over each open model at its recommended settings; speaker similarity (SIM) trails content accuracy, and Korean CER for Ming-3.0-tts-tiny (6.000%) is the clearest multilingual gap. The audio and ASR figures are read from images on the model cards; results marked with an asterisk there are self-evaluated. Competitor rows are not carried here. These results do not enter text-model or voice-agent rankings.

| Metric | Result |
| --- | --- |
| Seed-TTS testset Chinese WER | 0.820% |
| Seed-TTS testset English WER | 1.666% |
| Podcast CER | 2.03% |
| InstructTTSEval-ZH voice design average | 75.37% |
| CV3 emotion control average | 74.00% |

inclusionAI model cards retrieved October 21, 2026. Provider-reported; error rates are better lower. Voice results are display-only.

- [Evaluation source](https://huggingface.co/inclusionAI/Ming-3.0-tts-tiny)

MIT-licensed weights on Hugging Face and ModelScope, first uploaded September 22, 2026 and refreshed October 21, 2026. The card gives 7.9B total and about 1.3B active parameters; the checkpoint stores 8.98B including non-backbone modules. It serves with SGLang-Omni and supports the same seven languages and two dialects as the 0.5B model. The card states no context window.

## Other InclusionAI Models

- [Ling 3.1 Flash](/models/ling-3-1-flash) - Score: 68.05
- [Ling 3.0 Flash VL](/models/ling-3-0-flash-vl) - Score: 46.79
- [Ling 3.0 Flash](/models/ling-3-0-flash) - Score: 45.06
- [Ling 2.6 Flash](/models/ling-2-6-flash) - Score: 32.43
- [Ling 3.0 Flash Sante](/models/ling-3-0-flash-sante) - Score: not computed
- [Ling 3.0 Flash FP8](/models/ling-3-0-flash-fp8) - Score: not computed
- [Ling 3.0 Tiny](/models/ling-3-0-tiny) - Score: not computed
- [LLaDA2.2-flash](/models/llada2-2-flash) - Score: not computed
- [LLaDA2.2-mini](/models/llada2-2-mini) - Score: not computed
- [Ling 3.0 Flash Fin](/models/ling-3-0-flash-fin) - Score: not computed
- [Ming-3.0-asr-tiny](/models/ming-3-0-asr-tiny) - Score: not computed
- [Ming-3.0-audio-flash](/models/ming-3-0-audio-flash) - Score: not computed
- [Ming-3.0-audio-tiny](/models/ming-3-0-audio-tiny) - Score: not computed
- [Ming-3.0-tts-0.5B](/models/ming-3-0-tts-0-5b) - Score: not computed