Ling-3.0-tiny
- Reasoning
- Tool use
- Open weights
Prices
0 offerings · prices as listedNo priced offering is listed for this model yet.
Details
- Family
- ling-tiny
Not stated on Ant Group’s pages: max input, input modalities, output modalities, reasoning efforts, reasoning can be off, max thinking budget, structured output, built-in tools, knowledge cutoff, licence, aliases.
Run locally
Weights: inclusionAI/Ling-3.0-tiny- Parameters 7.9B
- Active 1.3B · 128 experts, 8 per token
- Native precision BF16
- Layers 24
- KV cache per token 27 KB
- Config positions 128K (config.json max_position_embeddings)
- API context 250K (maker’s API)
- Needs
- ≈ 15.6 GB (weights 14.7 (maker’s file) + KV cache 0.2 + overhead 0.7)
- Usable
- 21.6 GB (90% of its memory)
- Generation speed
- 178–267 tok/s (reads the active 1.3B per token)
- Longest context
- 128K (the model’s limit)
- First token
- about 0.04–0.06 s for a 2K prompt
| Precision | Weights | Needs at 8K | On your hardware | Longest context |
|---|---|---|---|---|
| BF16 (native)maker’s file | 14.7 GB | ≈ 15.6 | ●Runs well 178–267 tok/s | 128K |
| Q8_0maker’s file | 7.8 GB | ≈ 8.5 | ●Runs well 313–469 tok/s | 128K |
| FP8maker’s file | 7.8 GB | ≈ 8.5 | ●Runs well 313–469 tok/s | 128K |
| Q6_Kmaker’s file | 6.1 GB | ≈ 6.8 | ●Runs well 389–583 tok/s | 128K |
| Q5_K_Mmaker’s file | 5.2 GB | ≈ 6.0 | ●Runs well 437–655 tok/s | 128K |
| Q5_0 | 5.1 GB | ≈ 5.8 | ●Runs well 450–675 tok/s | 128K |
| Q4_K_Mmaker’s file | 4.5 GB | ≈ 5.2 | ●Runs well 494–740 tok/s | 128K |
| Q4_0 | 4.1 GB | ≈ 4.8 | ●Runs well 526–789 tok/s | 128K |
| INT4 (AWQ / GPTQ)maker’s file | 5.4 GB | ≈ 6.1 | ●Runs well 426–639 tok/s | 128K |
| IQ4_XS | 3.9 GB | ≈ 4.6 | ●Runs well 550–824 tok/s | 128K |
| Q3_K_Mquality drops | 3.6 GB | ≈ 4.3 | ●Runs well 585–877 tok/s | 128K |
| IQ3_XXSquality drops | 2.8 GB | ≈ 3.5 | ●Runs well 696–1045 tok/s | 128K |
| Q2_Kquality drops | 2.8 GB | ≈ 3.5 | ●Runs well 706–1059 tok/s | 128K |
| IQ2_XXSquality drops | 1.9 GB | ≈ 2.6 | ●Runs well 898–1347 tok/s | 128K |
Click a row to choose a precision. “Maker’s file” sizes are the maker’s own files; others are parameters × bits per weight.
Compared with other hardware
Each at the highest precision it runs. Click one for details.
Smallest that runs it (Q4_K_M, 8K): GPU GeForce RTX 3080; unified memory Apple M5 · 16 GB.
Run commands
vllm serve inclusionAI/Ling-3.0-tiny --max-model-len 8192
Memory to fine-tune
Batch 1, 2,048-token sequences, gradient checkpointing; LoRA and QLoRA train rank-16 adapters on the attention projections.
Estimates; facts from the maker’s Hugging Face repositories, as of 2026-11-06. How it’s estimated
Cited scores
From third-party leaderboards, not measured by usThis model isn’t on any leaderboard we cite yet.
Changes
All changes →No changes recorded yet.
Sources
- Introduction (opens in a new tab) huggingface.co · checked Nov 5, 2026
- Ling (opens in a new tab) developer.ant-ling.com · checked Oct 31, 2026
- Changelog (opens in a new tab) developer.ant-ling.com · checked Oct 31, 2026
- inclusionAI/Ling-3.0-tiny model card (opens in a new tab) huggingface.co · checked Oct 30, 2026
- inclusionAI/Ling-3.0-tiny config.json (opens in a new tab) huggingface.co · checked Nov 5, 2026
- inclusionAI/Ling-3.0-tiny files (opens in a new tab) huggingface.co · checked Nov 3, 2026
- inclusionAI/Ling-3.0-tiny-fp8 files (opens in a new tab) huggingface.co · checked Oct 31, 2026
- inclusionAI/Ling-3.0-tiny-GGUF files (opens in a new tab) huggingface.co · checked Nov 4, 2026
- inclusionAI/Ling-3.0-tiny-int4 files (opens in a new tab) huggingface.co · checked Oct 31, 2026
- inclusionAI/Ling-3.0-tiny README.md (opens in a new tab) huggingface.co · checked Nov 3, 2026
Some values on this page are still waiting to be checked against official pages: a filled dot is verified, a hollow one is not yet (with or without a source on file). How we verify data
Ant Group · China · developer.ant-ling.com
Related models
Snapshots, the same family and other models from this provider