CodeQwen1.5 7B — Hardware Requirements & GPU Compatibility
ChatCode
CodeQwen1.5-7B is Alibaba's 7.3-billion-parameter code-specialized base language model, built on the Qwen1.5 architecture and pretrained on 3 trillion tokens of code data covering 92 programming languages. It uses group-query attention for efficient inference and is a raw pretrained model, not tuned for chat; Alibaba advises against using it directly for conversation and instead for fine-tuning, code infilling, and code generation. The card highlights strong text-to-SQL and bug-fixing performance for its size. At 7B parameters it fits comfortably on a single consumer GPU, even less once quantized. Context length is 65,536 tokens. It is released under a custom Tongyi Qianwen Research License, free for academic research, with commercial use requiring direct contact with Alibaba. It was published in April 2024.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 32K)Best on 24 GB (Q4_K_M · ~137 tok/s · 64K)Macs from 8 GB (Q2_K · ~20 tok/s)
Other
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run CodeQwen1.5 7B
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
ollama run hf.co/mradermacher/CodeQwen1.5-7B-i1-GGUF:Q4_K_MQ4_K_M fits fully on the RTX 4090 · 18.3 GB free · 16K context
Download: CodeQwen1.5-7B.i1-Q4_K_M.gguf4.74 GBmradermacher
Ways to runOther ways
each opens step-by-stepRun in cloud
Your own private copy · Q4_K_M · 24 GB GPU
- The easiest way to try this model
- Running in a few minutes, easy setup
- Chat right here or connect any app
$5+ bonus for new users
Runs on RTX 3080 Ti (12 GB) (6 GB headroom) · Q4_K_M
- Generation speed
- ~124 tok/s
- generation speed
- Cost per 1M output tokens
- $0.27
- per 1M output tokens
How Much VRAM Does CodeQwen1.5 7B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 3.5 GB | 5.5 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 64K context · ~186 tok/s | |
| 3.6 GB | 5.6 GB | 3-bit small — 3-bit small quantization | Fits: 64K context · ~182 tok/s | |
| 4.0 GB | 6 GB | 3-bit medium — 3-bit medium quantization | Fits: 64K context · ~165 tok/s | |
| 4.1 GB | 6.1 GB | 4-bit · legacy — 4-bit legacy quantization | Fits: 64K context · ~161 tok/s | |
| 4.8 GB | 6.8 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 64K context · ~137 tok/s | |
| 5.6 GB | 7.6 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 64K context · ~117 tok/s | |
| 6.4 GB | 8.4 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 64K context · ~102 tok/s | |
| 7.7 GB | 9.7 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 64K context · ~85 tok/s |
Selected Q4_K_M
4-bit medium quantization5.7GBat 16K context
- Weights
- 4.4 GB
- KV cache
- 1.1 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 18.3 GB free · ~137 tok/s
≥50 tok/s needs ≥367 GB/s · RTX 5090 ~244, RTX 3090 Ti ~137
Which Hardware Can Run CodeQwen1.5 7B?
Q4_K_M · 4.8 GBCodeQwen1.5 7B (Q4_K_M) requires 4.8 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 7+ GB is recommended. Using the full 66K context window can add up to 4.2 GB, bringing total usage to 8.9 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 151 devices with unified memory can run CodeQwen1.5 7B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).
≤ 11 GB
≤ 11 GB tier →fits · 2.3 GB free · up to 32K ctx
16 GPUs · 2 device families
RTX 3080~103Simulate output on RTX 3080RTX 2080 Ti~84Simulate output on RTX 2080 TiRTX 3070 Ti~83Simulate output on RTX 3070 TiOrin Nano~14Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~10Simulate output on iPhone 1720 GB
fits · 14.3 GB free · up to 64K ctx
1 GPU
RX 7900 XT~100Simulate output on RX 7900 XT32–40 GB
fits · 26.3 GB free · up to 64K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~244Simulate output on RTX 5090M5 Max36 GB~67Simulate output on Mac Studio M5 Max (36 GB)M532 GB~23Simulate output on MacBook Air 13" M5 (32 GB)Orin~28Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~18Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
How good is CodeQwen1.5 7B?
llmrun Score blends public benchmark boards into one number; 100 = the average open model.
Not enough data yet for an llmrun Score.
A Score needs results on enough independent boards. How the Composite Score is built
Benchmarks
Here are all 4 boards that have results for this model, sorted by their weight in the llmrun Score; badges show who ran the test.
- MMLU
Epoch AI · Apr 2024Lab-reported
Lab-reported40.5%
#70/85 open
- GSM8K
Epoch AI · Apr 2024Lab-reported
Lab-reported37.7%
#37/62 open
- ARC Challenge
Epoch AI · Apr 2024Lab-reported
Lab-reported35.7%
#49/56 open
- WinoGrande
Epoch AI · Apr 2024Lab-reported
Lab-reported59.8%
#43/51 open
Alternatives to CodeQwen1.5 7B
Qwen family · 28 models →Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Olmo 3 7B InstructAllen AI | — | 8.4 GB | ~114 | same size |
| Phi Mini MoE InstructMicrosoft | — | 5.4 GB | ~239 | same size |
| Falcon H1 7B InstructTII UAE | — | 6.3 GB | ~130 | same size |
| DeepSeek R1 Distill Qwen 7BDeepSeek | — | 5.8 GB | ~131 | same size |
| Eurus 2 7B PRIMEPRIME-RL | — | 5.1 GB | ~131 | same size |
| Mathstral 7B v0.1Mistral AI | — | 6.8 GB | ~133 | same size |
Specifications
- Publisher
- Alibaba
- Family
- Qwen
- Parameters
- 7.3B
- Architecture
- Qwen2ForCausalLM
- Context Length
- 65,536 tokens
- Vocabulary Size
- 92,416
- Release Date
- 2024-04-15
- License
- Other
Frequently Asked Questions
How much VRAM does CodeQwen1.5 7B need?
CodeQwen1.5 7B requires 4.8 GB of VRAM at Q4_K_M, or 14.9 GB at BF16. Full 66K context adds up to 4.2 GB (8.9 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 7.3B × 4.8 bits ÷ 8 = 4.4 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 4.5 GB (at full 66K context)
VRAM usage by quantization
What's the best quantization for CodeQwen1.5 7B?
For CodeQwen1.5 7B, Q4_K_M (4.8 GB) offers the best balance of quality and VRAM usage. Q5_K_S (5.4 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 2.4 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run CodeQwen1.5 7B on a Mac?
Yes — Mac Mini M2 (8 GB) and 145 other Macs can run CodeQwen1.5 7B, the smallest at Q2_K. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 4.8 GB of usable unified memory (RAM minus macOS overhead), from Mac Mini M2 Pro (16 GB).
Can I run CodeQwen1.5 7B locally?
Yes — CodeQwen1.5 7B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is CodeQwen1.5 7B?
At Q4_K_M, CodeQwen1.5 7B can reach ~1004 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~137 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 4.78 × 0.65 = ~1088 tok/s
Estimated speed at Q4_K_M (4.8 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of CodeQwen1.5 7B?
At Q4_K_M, the download is about 4.35 GB. The full-precision BF16 version is 14.50 GB. The smallest option (IQ2_XXS) is 1.99 GB.
Which GPUs can run CodeQwen1.5 7B?
52 consumer GPUs can run CodeQwen1.5 7B at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.
Which devices can run CodeQwen1.5 7B?
151 devices with unified memory can run CodeQwen1.5 7B at Q4_K_M (4.8 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.