Ant Group models, access and benchmark guide
Ant Group model and access guide with a practical US-English evaluation brief and links to publisher materials.

👁 111 reads
The inclusionAI repositories span language and speech. A text benchmark for Ling gives no verdict on a voice generated by Ming. I would identify the exact checkpoint, license and input contract before assigning either one a role.
Choosing the first job
Extract an address with Ling, then synthesize the checked address with Ming and listen for the apartment number.
Models in the November 17 snapshot
| Model and track | Placement | Reported score |
|---|---|---|
| Ming-3.0-tts (AA provider voice) | 24 | 1182 Elo |
| Ming-3.0-tts (AA controlled voice) | 18 | 1070 Elo |
| Ling 3.1 Flash 1019 (AA intelligence) | 22 | 30 index points |
| Ling 3.1 Flash 1019 (AA additions) |
Release record. Publisher documentation identifies the product family. The ranked configurations include Ming-3.0-tts, Ling 3.1 Flash 1019.
Google · Your sourcesQuantslant higher in your Google - one clickAdd to preferred sources →Sources
- https://huggingface.co/inclusionAI (accessed November 17, 2026)
- https://github.com/inclusionAI (accessed November 17, 2026)
- https://artificialanalysis.ai/text-to-speech/leaderboard/provider-voice (accessed November 17, 2026)
- https://artificialanalysis.ai/text-to-speech/leaderboard/controlled-voice (accessed November 17, 2026)
- https://artificialanalysis.ai/leaderboards/models (accessed November 17, 2026)
- https://artificialanalysis.ai/changelog (accessed November 17, 2026)
More on this from QUANTSLANT
benchmarks·Amazon models, access and benchmark guideAmazon model and access guide with a practical US-English evaluation brief and links to publisher materials.