// Qwen · Vision-language
Qwen3.5-122B-A10B: VRAM & local deployment
125.1B total · 10B active (MoE) parameters. About 84.1 GB at 4-bit and 276 GB at BF16 with one 8K-token request.
released 2026-02-24licence: Commercial OKmemory figures are estimates
01Specs
| Publisher | Alibaba Qwen · Qwen |
|---|---|
| Family / series | Qwen · Qwen3.5 |
| Type | Vision-language (image-text-to-text) |
| Parameters | 125.1B total · 10B active (MoE) Source: counted from official safetensors weights |
| Context length | 256K tokens (262,144)max_position_embeddings in config.json |
| Architecture | 48 layers (12 full-attention) · GQA+HYBRID · MoE |
| Released weights | BF16 · 250 GBdownload |
| Licence | Apache-2.0 · Commercial OKPermissive licence (Apache-2.0 / MIT / CC-BY / OpenMDW): commercial use allowed. |
| Released | 2026-02-24 (Hugging Face repo created) |
02How much VRAM it needs (estimate)
Total = weights + 10% + 0.5GB overhead + KV cache for one request (FP16 KV). More concurrent users multiply the KV column.
| Precision | Weights | Total @ 8K | Total @ 32K | Total @ 128K | Smallest GPU | Mac |
|---|---|---|---|---|---|---|
| BF16 / FP16 (original) | 250 GB | 276 GB | 276 GB | 279 GB | 4 × RTX PRO 6000 (384GB) | Mac 512GB |
| FP8 / INT8 / Q8_0 | 133 GB | 147 GB | 148 GB | 150 GB | 4 × RTX PRO 6000 (384GB) | Mac 256GB |
| 4-bit (Q4_K_M / AWQ / GPTQ-Int4) | 75.8 GB | 84.1 GB | 84.7 GB | 87.1 GB | RTX PRO 6000 96GB | Mac 128GB |
KV cache per request: 8K 0.2 GB · 32K 0.8 GB · 128K 3.2 GB
Bits per weight: BF16 16, FP8 8.5, 4-bit 4.85. “Fits” = total ≤ 90% of memory; Mac GPU can use ~2/3 of unified memory (≤36GB) or ~3/4 (larger). For MoE, all experts must be in memory, so size by total parameters. Method →
03How to run it
Only runtimes the publisher or the runtime’s official library names for this model.
- vLLM — mentioned in the official model card
- SGLang — mentioned in the official model card
- Transformers — mentioned in the official model card
- KTransformers — mentioned in the official model card
- Ollama — in the official Ollama library · qwen3.5:122b-a10b
ollama run qwen3.5:122b-a10b04Official repositories & quantised builds
3 official repo(s) from Qwen.
- Qwen3.5-122B-A10B main
- Qwen3.5-122B-A10B-GPTQ-Int4 GPTQ/INT4
- Qwen3.5-122B-A10B-FP8 FP8
05Other Qwen models
All 112 →| Model | Params | 4-bit total | Released |
|---|---|---|---|
| Qwen3.5-0.8B · same series | 873M | 1.2 GB | 2026-02-28 |
| Qwen3.5-2B · same series | 2.27B | 2.1 GB | 2026-02-28 |
| Qwen3.5-4B · same series | 4.66B | 3.9 GB | 2026-02-27 |
| Qwen3.5-9B · same series | 9.65B | 7.2 GB | 2026-02-27 |
| Qwen3.5-27B · same series | 27.8B | 19.6 GB | 2026-02-24 |
| Qwen3.5-35B-A3B · same series | 36B / 3B | 24.6 GB | 2026-02-24 |
| Qwen3.5-397B-A17B · same series | 403.4B / 17B | 270 GB | 2026-02-16 |
| Qwen-Image-2.1-Turbo | 7.12B | 5.2 GB | 2026-10-09 |
| Qwen-Image-2.1 | 7.12B | 5.2 GB | 2026-09-14 |
| Qwen3.8-Flash-Next | 180B / 6B | 121 GB | 2026-08-24 |
| Qwen3.8-2.4T-A95B | 2.4T / 95B | 1.60 TB | 2026-08-08 |
| Qwen3.8-27B | 27.8B | 19.6 GB | 2026-08-05 |
| Qwen3.6-27B | 27.8B | 19.6 GB | 2026-04-21 |
| Qwen3.6-35B-A3B | 36B / 3B | 24.6 GB | 2026-04-15 |
| Qwen3-Coder-Next | 80B / 3B | 54.1 GB | 2026-01-30 |
06Sources
Data checked 2026-04-24. Parameters, context and licence are copied from the official repository; memory figures are our estimates.