Ling-3.0-flash-VL
- In Input modalities: TextImageVideo Out Output modalities: Text
- Reasoning
- Tool use
- Open weights
Prices
1 offering · prices as listedLing-3.0-flash-VL| Conditions/ 1M tokens | Input | Cache read | Output |
|---|---|---|---|
| Standard | ¥0.50 | ¥0.10 | ¥1.50 |
Prices from Ant Ling: model pricing (opens in a new tab) (checked Nov 4, 2026)
Standard: the price a request with no conditions pays (the headline price elsewhere on the site).
Details
- Licence
- MIT
- Family
- ling-flash-vl
- Model ID
Not stated on Ant Group’s pages: max input, reasoning efforts, reasoning can be off, max thinking budget, structured output, built-in tools, knowledge cutoff, aliases.
Run locally
Weights: inclusionAI/Ling-3.0-flash-VL- Parameters 124.8B
- Active 5.5B · 512 experts, 8 per token
- Native precision BF16
- Layers 42
- KV cache per token 47 KB
- Config positions 128K (config.json max_position_embeddings)
- API context 250K (maker’s API)
- Needs
- ≈ 74.5 GB (weights 70.6 + KV cache 0.4 + overhead 3.5)
- Usable
- 21.6 GB (90% of its memory)
- In system memory
- 53.8 GB · 10/42 layers on the GPU (--n-gpu-layers 10)
- Generation speed
- 17–26 tok/s (reads the active 5.5B per token)
- First token
- about 0.18–0.27 s for a 2K prompt
| Precision | Weights | Needs at 8K | On your hardware | Longest context |
|---|---|---|---|---|
| BF16 (native)maker’s file | 232.6 GB | ≈ 244.6 | ✕Doesn’t fit | — |
| Q8_0 | 123.5 GB | ≈ 130.1 | ✕Doesn’t fit | — |
| FP8maker’s file | 117.5 GB | ≈ 123.8 | ✕Doesn’t fit | — |
| Q6_K | 95.5 GB | ≈ 100.6 | ✕Doesn’t fit | — |
| Q5_K_M | 82.7 GB | ≈ 87.2 | ✕Doesn’t fit | — |
| Q5_0 | 79.9 GB | ≈ 84.3 | ✕Doesn’t fit | — |
| Q4_K_M | 70.6 GB | ≈ 74.5 | ○Needs system memory 17–26 tok/s | — |
| Q4_0 | 65.4 GB | ≈ 69.0 | ○Needs system memory 19–29 tok/s | — |
| INT4 (AWQ / GPTQ)maker’s file | 70.9 GB | ≈ 74.8 | ○Needs system memory 17–26 tok/s | — |
| IQ4_XS | 61.8 GB | ≈ 65.2 | ○Needs system memory 21–31 tok/s | — |
| Q3_K_Mquality drops | 56.8 GB | ≈ 60.0 | ○Needs system memory 41–62 tok/s | — |
| IQ3_XXSquality drops | 44.5 GB | ≈ 47.1 | ○Needs system memory 53–79 tok/s | — |
| Q2_Kquality drops | 43.6 GB | ≈ 46.2 | ○Needs system memory 54–81 tok/s | — |
| IQ2_XXSquality drops | 29.9 GB | ≈ 31.8 | ○Needs system memory 79–118 tok/s | — |
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 NVIDIA RTX PRO 6000 Blackwell Workstation Edition; unified memory Apple M5 Max (40-core GPU) · 128 GB.
Run commands
The maker publishes no files at this precision. Choose its native precision or one of its own builds, or quantise it yourself with llama.cpp.
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.
News
All news →- Sep 4, 2026Release Ant Group releases Ling-3.0-flash-VL On September 4, 2026, Ant Group launched multimodal Ling-3.0-flash-VL. It can be tried in the chat UI and called through OpenAI-compatible and Anthropic-compatible APIs.
Price history
Ant Ling API, China · standard rate · as listedInput per 1M tokens
| From | Price |
|---|---|
| Sep 28, 2026 | ¥0.14 |
| Oct 28, 2026 | ¥0.50 |
Output per 1M tokens
| From | Price |
|---|---|
| Sep 28, 2026 | ¥0.42 |
| Oct 28, 2026 | ¥1.50 |
Changes
All changes →- New language model
Sources
- Ant Ling: model pricing (opens in a new tab) developer.ant-ling.com · checked Nov 4, 2026
- Ling (opens in a new tab) developer.ant-ling.com · checked Nov 1, 2026
- Changelog (opens in a new tab) developer.ant-ling.com · checked Nov 1, 2026
- inclusionAI/Ling-3.0-flash-VL model card (opens in a new tab) huggingface.co · checked Nov 1, 2026
- inclusionAI/Ling-3.0-flash-VL README.md (opens in a new tab) huggingface.co · checked Nov 1, 2026
- models.dev: inclusionai/ling-3.0-flash-vl (community data, MIT) (opens in a new tab) github.com · checked Oct 16, 2026
- inclusionAI/Ling-3.0-flash-VL model card (opens in a new tab) huggingface.co · checked Nov 1, 2026
- inclusionAI/Ling-3.0-flash-VL model card (opens in a new tab) huggingface.co · checked Oct 30, 2026
- inclusionAI/Ling-3.0-flash-VL config.json (opens in a new tab) huggingface.co · checked Oct 30, 2026
- inclusionAI/Ling-3.0-flash-VL files (opens in a new tab) huggingface.co · checked Oct 31, 2026
- inclusionAI/Ling-3.0-flash-VL-fp8 files (opens in a new tab) huggingface.co · checked Oct 31, 2026
- inclusionAI/Ling-3.0-flash-VL-int4 files (opens in a new tab) huggingface.co · checked Oct 30, 2026
- inclusionAI/Ling-3.0-flash-VL README.md (opens in a new tab) huggingface.co · checked Oct 30, 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