Qwen3.5 4B hardware requirements
What is Qwen3.5 4B?
Qwen3.5 4B is one of the small Qwen3.5 models that the Qwen team released in March 2026. It reads images and video as well as text, and it thinks before answering by default unless enable_thinking is set to false. Its model card highlights coding, agents, tool calling and 201 languages and dialects; the Apache 2.0 license allows commercial use. Only one layer in four uses full attention, which shrinks its KV cache to about a quarter.
About the model
- Parameters
- 4.66B
- Attention
- Hybrid attention (only some layers use full attention)
- Max context
- 262,144 tokens
- Released
- 2026-02-27
- Model card
- Qwen3.5 4B on Hugging Face
Other sizes in this series
| Model | Parameters | Baseline quant | Memory needed (8k) | GeForce RTX 4060 | GeForce RTX 5060 Ti 16GB | GeForce RTX 3090 |
|---|---|---|---|---|---|---|
| Qwen3.5 4Bthis model | 4.66B | Q4_K_M | 3.6 GB | Runs great | Runs great | Runs great |
| Qwen3.5 9B | 9.65B | Q4_K_M | 6.0 GB | Runs great | Runs great | Runs great |
| Qwen3.5 27B | 27.78B | Q4_K_M | 17.2 GB | Runs slowly | Runs slowly | Runs great |
| Qwen3.5 35B-A3B | 35.95B total · 3B active (MoE) | Q4_K_M | 22.3 GB | Runs well | Runs well | Runs great |
| Qwen3.5 122B-A10B | 125.09B total · 10B active (MoE) | Q4_K_M | 76.6 GB | Won't run | Won't run | Won't run |
GGUF files
| Quant | File size | Quality | Published by |
|---|---|---|---|
| UD-Q2_K_XL | 2.12 GB | Heavy loss | Community GGUF · unsloth |
| IQ4_XS | 2.64 GB | Good balance | Community GGUF · unsloth |
| Q4_K_Mbaseline | 2.83 GB | Good balance | Community GGUF · unsloth |
| Q8_0 | 4.61 GB | Near-lossless | Community GGUF · unsloth |
Some repositories publish only Unsloth Dynamic (UD) variants; sizes are those of the file you would download.
Memory needed by quant and context
| Quant | 4k | 8k | 16k | 32k | 64k | 128k | 256k |
|---|---|---|---|---|---|---|---|
| UD-Q2_K_XL | 2.7 | 2.9 | 3.2 | 3.9 | 5.3 | 8.1 | 13.5 |
| IQ4_XS | 3.2 | 3.4 | 3.7 | 4.4 | 5.8 | 8.5 | 14.0 |
| Q4_K_M | 3.4 | 3.6 | 3.9 | 4.6 | 6.0 | 8.7 | 14.2 |
| Q8_0 | 5.1 | 5.3 | 5.7 | 6.3 | 7.7 | 10.5 | 16.0 |
Which GPUs can run Qwen3.5 4B?
How to read the verdicts
- Runs great
- Fully on the GPU at 20 tok/s or more
- Runs well
- Fully on the GPU at 8–20 tok/s; MoE experts in system RAM at 20 tok/s or more; or at least 90% on the GPU at 8 tok/s or more
- Runs slowly
- 2–8 tok/s; CPU-only; less than 90% on the GPU; or MoE experts in system RAM below 20 tok/s
- Won't run
- Does not fit, or under 2 tok/s
Filter this list by budget, type and context in the calculator
Scroll sideways to see every column.
| Hardware | Verdict | Speed | Quant | Memory | Runs as | Context | Notes | Price |
|---|---|---|---|---|---|---|---|---|
| GeForce RTX 5090 | Runs great | est. 404.8 tok/s356.3–453.4calibrated estimate ±12% | Q4_K_M | 4.2 / 32.0 GB | Full GPU | up to 256k | — | $4,200 street (as of 2026-08-09) |
| Radeon RX 7900 XTX | Runs great | est. 235.5 tok/s207.2–263.7calibrated estimate ±12% | Q4_K_M | 4.2 / 24.0 GB | Full GPU | up to 256k | — | $999 launch MSRP |
| GeForce RTX 3090 Ti | Runs great | est. 227.7 tok/s200.4–255.1calibrated estimate ±12% | Q4_K_M | 4.2 / 24.0 GB | Full GPU | up to 256k | — | $1,999 launch MSRP |
| GeForce RTX 4090 | Runs great | est. 227.7 tok/s200.4–255.1calibrated estimate ±12% | Q4_K_M | 4.2 / 24.0 GB | Full GPU | up to 256k | — | $2,755 street (as of 2026-08-09) |
| Apple M5 Ultra · 96 GB | Runs great | est. 225.9 tok/s180.7–271.1calibrated estimate ±20% | Q4_K_M | 3.7 / 67.2 GB | Unified memory | up to 256k | — | — |
| GeForce RTX 5080 | Runs great | est. 216.9 tok/s190.9–242.9calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 256k | — | $1,256 street (as of 2026-08-09) |
| GeForce RTX 3090 | Runs great | est. 211.5 tok/s186.1–236.8calibrated estimate ±12% | Q4_K_M | 4.2 / 24.0 GB | Full GPU | up to 256k | — | $1,050 street (as of 2026-08-09) |
| GeForce RTX 3080 12GB | Runs great | est. 206.0 tok/s181.3–230.8calibrated estimate ±12% | Q4_K_M | 4.2 / 12.0 GB | Full GPU | up to 128k | — | $799 launch MSRP |
| GeForce RTX 5070 Ti | Runs great | est. 202.4 tok/s178.1–226.7calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 256k | — | $949 street (as of 2026-08-09) |
| GeForce RTX 5080 Laptop | Runs great | est. 202.4 tok/s178.1–226.7calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 256k | — | — |
| GeForce RTX 5090 Laptop | Runs great | est. 202.4 tok/s178.1–226.7calibrated estimate ±12% | Q4_K_M | 4.2 / 24.0 GB | Full GPU | up to 256k | — | — |
| Radeon RX 7900 XT | Runs great | est. 196.2 tok/s172.7–219.8calibrated estimate ±12% | Q4_K_M | 4.2 / 20.0 GB | Full GPU | up to 256k | — | $899 launch MSRP |
| GeForce RTX 3080 10GB | Runs great | est. 171.7 tok/s151.1–192.3calibrated estimate ±12% | Q4_K_M | 4.2 / 10.0 GB | Full GPU | up to 128k | — | $699 launch MSRP |
| GeForce RTX 4080 Super | Runs great | est. 166.3 tok/s146.3–186.2calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 256k | — | $999 launch MSRP |
| GeForce RTX 4080 | Runs great | est. 162.0 tok/s142.5–181.4calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 256k | — | $1,199 launch MSRP |
| Radeon RX 9070 | Runs great | est. 158.2 tok/s139.2–177.2calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 256k | — | $639 street (as of 2026-08-09) |
| Radeon RX 9070 XT | Runs great | est. 158.2 tok/s139.2–177.2calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 256k | — | $689 street (as of 2026-08-09) |
| Radeon RX 7800 XT | Runs great | est. 153.1 tok/s134.7–171.4calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 256k | — | $499 launch MSRP |
| GeForce RTX 4070 Ti Super | Runs great | est. 151.8 tok/s133.6–170.0calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 256k | — | $799 launch MSRP |
| GeForce RTX 5070 | Runs great | est. 151.8 tok/s133.6–170.0calibrated estimate ±12% | Q4_K_M | 4.2 / 12.0 GB | Full GPU | up to 128k | — | $629 street (as of 2026-08-09) |
| GeForce RTX 5070 Ti Laptop | Runs great | est. 151.8 tok/s133.6–170.0calibrated estimate ±12% | Q4_K_M | 4.2 / 12.0 GB | Full GPU | up to 128k | — | — |
| Apple M3 Ultra · 96 GB | Runs great | est. 150.7 tok/s120.5–180.8calibrated estimate ±20% | Q4_K_M | 3.7 / 67.2 GB | Unified memory | up to 256k | — | — |
| Apple M5 Max (40-core GPU) · 48 GB | Runs great | est. 150.6 tok/s120.5–180.7calibrated estimate ±20% | Q4_K_M | 3.7 / 33.6 GB | Unified memory | up to 256k | — | — |
| Apple M2 Ultra · 64 GB | Runs great | est. 147.2 tok/s117.7–176.6calibrated estimate ±20% | Q4_K_M | 3.7 / 44.8 GB | Unified memory | up to 256k | — | — |
| GeForce RTX 2080 Ti | Runs great | est. 139.2 tok/s122.5–155.9calibrated estimate ±12% | Q4_K_M | 4.2 / 11.0 GB | Full GPU | up to 128k | — | $999 launch MSRP |
| Apple M4 Max (40-core GPU) · 48 GB | Runs great | est. 133.9 tok/s107.1–160.7calibrated estimate ±20% | Q4_K_M | 3.7 / 33.6 GB | Unified memory | up to 256k | — | — |
| GeForce RTX 4090 Laptop | Runs great | est. 130.1 tok/s114.5–145.7calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 256k | — | — |
| GeForce RTX 4070 | Runs great | est. 113.9 tok/s100.2–127.5calibrated estimate ±12% | Q4_K_M | 4.2 / 12.0 GB | Full GPU | up to 128k | — | $599 launch MSRP |
| GeForce RTX 4070 Super | Runs great | est. 113.9 tok/s100.2–127.5calibrated estimate ±12% | Q4_K_M | 4.2 / 12.0 GB | Full GPU | up to 128k | — | $599 launch MSRP |
| Apple M5 Max (32-core GPU) · 36 GB | Runs great | est. 112.8 tok/s90.3–135.4calibrated estimate ±20% | Q4_K_M | 3.7 / 25.2 GB | Unified memory | up to 256k | — | — |
| GeForce GTX 1080 Ti | Runs great | est. 109.3 tok/s76.5–142.1theoretical estimate ±30% | Q4_K_M | 4.2 / 11.0 GB | Full GPU | up to 128k | — | $699 launch MSRP |
| GeForce RTX 3070 | Runs great | est. 101.2 tok/s89.1–113.4calibrated estimate ±12% | Q4_K_M | 4.2 / 8.0 GB | Full GPU | up to 64k | — | $499 launch MSRP |
| GeForce RTX 5060 | Runs great | est. 101.2 tok/s89.1–113.4calibrated estimate ±12% | Q4_K_M | 4.2 / 8.0 GB | Full GPU | up to 64k | — | $339 street (as of 2026-08-09) |
| GeForce RTX 5060 Ti 16GB | Runs great | est. 101.2 tok/s89.1–113.4calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 128k | — | $569 street (as of 2026-08-09) |
| GeForce RTX 5060 Ti 8GB | Runs great | est. 101.2 tok/s89.1–113.4calibrated estimate ±12% | Q4_K_M | 4.2 / 8.0 GB | Full GPU | up to 64k | — | $429 street (as of 2026-08-09) |
| Apple M4 Max (32-core GPU) · 36 GB | Runs great | est. 100.6 tok/s80.5–120.7calibrated estimate ±20% | Q4_K_M | 3.7 / 25.2 GB | Unified memory | up to 128k | — | — |
| GeForce RTX 4080 Laptop | Runs great | est. 97.6 tok/s85.9–109.3calibrated estimate ±12% | Q4_K_M | 4.2 / 12.0 GB | Full GPU | up to 128k | — | — |
| Intel Arc B580 | Runs great | est. 86.8 tok/s69.5–104.2calibrated estimate ±20% | Q4_K_M | 4.2 / 12.0 GB | Full GPU | up to 128k | — | $290 street (as of 2026-08-09) |
| GeForce RTX 3060 12GB | Runs great | est. 81.3 tok/s71.6–91.1calibrated estimate ±12% | Q4_K_M | 4.2 / 12.0 GB | Full GPU | up to 128k | — | $250 street (as of 2026-08-09) |
| Radeon RX 9060 XT 16GB | Runs great | est. 79.0 tok/s69.5–88.5calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 128k | — | $449 street (as of 2026-08-09) |
| Apple M5 Pro · 24 GB | Runs great | est. 75.3 tok/s60.2–90.4calibrated estimate ±20% | Q4_K_M | 3.7 / 16.8 GB | Unified memory | up to 128k | — | — |
| NVIDIA DGX Spark · 64 GB | Runs great | est. 68.7 tok/s55.0–82.5calibrated estimate ±20% | Q4_K_M | 3.7 / 48.2 GB | Unified memory | up to 64k | — | $3,999 launch MSRP (128 GB) |
| Ryzen AI Max+ 395 (Strix Halo) · 32 GB | Runs great | est. 67.8 tok/s54.2–81.3calibrated estimate ±20% | Q4_K_M | 3.7 / 22.4 GB | Unified memory | up to 64k | — | $1,999 launch MSRP (128 GB) |
| Apple M4 Pro · 24 GB | Runs great | est. 67.0 tok/s53.6–80.4calibrated estimate ±20% | Q4_K_M | 3.7 / 16.8 GB | Unified memory | up to 64k | — | — |
| GeForce RTX 4060 Ti 16GB | Runs great | est. 65.1 tok/s57.3–72.9calibrated estimate ±12% | Q4_K_M | 4.2 / 16.0 GB | Full GPU | up to 64k | — | $499 launch MSRP |
| GeForce RTX 4060 Ti 8GB | Runs great | est. 65.1 tok/s57.3–72.9calibrated estimate ±12% | Q4_K_M | 4.2 / 8.0 GB | Full GPU | up to 64k | — | $399 launch MSRP |
| GeForce RTX 4060 | Runs great | est. 61.5 tok/s54.1–68.8calibrated estimate ±12% | Q4_K_M | 4.2 / 8.0 GB | Full GPU | up to 64k | — | $299 launch MSRP |
| GeForce RTX 4060 Laptop | Runs great | est. 57.8 tok/s50.9–64.8calibrated estimate ±12% | Q4_K_M | 4.2 / 8.0 GB | Full GPU | up to 64k | — | — |
| GeForce RTX 4070 Laptop | Runs great | est. 57.8 tok/s50.9–64.8calibrated estimate ±12% | Q4_K_M | 4.2 / 8.0 GB | Full GPU | up to 64k | — | — |
| GeForce RTX 3050 8GB | Runs great | est. 50.6 tok/s44.5–56.7calibrated estimate ±12% | Q4_K_M | 4.2 / 8.0 GB | Full GPU | up to 64k | — | $249 launch MSRP |
| Apple M5 · 16 GB | Runs great | est. 37.7 tok/s30.1–45.2calibrated estimate ±20% | Q4_K_M | 3.7 / 11.2 GB | Unified memory | up to 16k | — | — |
| Apple M6 · 16 GB | Runs great | est. 37.7 tok/s30.1–45.2calibrated estimate ±20% | Q4_K_M | 3.7 / 11.2 GB | Unified memory | up to 16k | — | — |
| Apple M4 · 16 GB | Runs well | est. 29.4 tok/s23.5–35.3calibrated estimate ±20% | Q4_K_M | 3.7 / 11.2 GB | Unified memory | up to 128k |
| — |
| DDR5-6000 dual-channel CPU · 16 GB | Runs slowly | est. 15.5 tok/s12.4–18.6calibrated estimate ±20% | Q4_K_M | 3.1 GB RAM | CPU only | up to 256k |
| — |
| DDR5-5600 dual-channel CPU · 16 GB | Runs slowly | est. 14.5 tok/s11.6–17.4calibrated estimate ±20% | Q4_K_M | 3.1 GB RAM | CPU only | up to 256k |
| — |
| DDR4-3200 dual-channel CPU · 16 GB | Runs slowly | est. 8.3 tok/s6.6–9.9calibrated estimate ±20% | Q4_K_M | 3.1 GB RAM | CPU only | up to 128k | — | — |
Models with a thinking mode spend extra tokens thinking, so their speed thresholds are 1.5× stricter (30 / 12 / 3 tok/s). With thinking turned off, a model may do one grade better.
Cheapest GPUs that run it
- Runs greatGeForce RTX 3060 12GB$250 street (as of 2026-08-09), est. 81.3 tok/s (71.6–91.1, calibrated estimate ±12%).
Measured results for Qwen3.5 4B
No public measurements for this model yet.
Frequently asked questions
How much VRAM does Qwen3.5 4B need?
At Q4_K_M the file is 2.83 GB; with 8k context a graphics card needs about 3.6 GB (weights 2.74 GB, KV cache 0.3 GB, compute buffer 0.6 GB). The smallest tracked file (UD-Q2_K_XL (heavy quality loss)) needs about 2.9 GB. Part of the file stays off the graphics card: llama.cpp keeps the input embedding table in system RAM and does not use this model’s MTP (multi-token prediction) block yet, so it never loads it. A Mac counts the whole file, so it needs a little more.
What is the cheapest GPU that runs Qwen3.5 4B well?
Cheapest that runs it great: GeForce RTX 3060 12GB at $250 (street price as of 2026-08-09), est. 81.3 tok/s (71.6–91.1, calibrated estimate ±12%).
Can Qwen3.5 4B run on an 8 GB, 16 GB or 24 GB GPU?
At Q4_K_M with 8k context: GeForce RTX 4060 — Runs great, est. 61.5 tok/s (54.1–68.8, calibrated estimate ±12%); GeForce RTX 5060 Ti 16GB — Runs great, est. 101.2 tok/s (89.1–113.4, calibrated estimate ±12%); GeForce RTX 3090 — Runs great, est. 211.5 tok/s (186.1–236.8, calibrated estimate ±12%).
Speeds are estimates from memory bandwidth, calibrated against public benchmarks, and each one comes with an error band and a confidence label. Real results vary with drivers, backend, context length and thermals.