How much VRAM does WizardCoder 33B v1.1 need?
WizardCoder 33B v1.1 needs about 20.5 GB of VRAM at Q4_K_M, or 35.3 GB at Q8_0 and 65.9 GB at FP16. That fits a consumer card: the smallest in our set that holds it at Q4_K_M is the RTX 3090 (24 GB). Weights plus runtime overhead; with only 8K of context the cache stays small.
Weights-plus-runtime footprint across the six most common quantizations. KV cache is context-dependent and comes on top. A full 8K window is a real share of the budget; size it precisely in the GPU Memory Calculator.
Can your GPU run WizardCoder 33B v1.1?
Pick your card — the answer below is computed by the same engine that produces every figure on this page.
RTX 3090 runs WizardCoder 33B v1.1 at Q4_K_M with 2.3 GB to spare; room left for context and everyday runtime pressure.
Weights plus runtime overhead. KV cache is not included — it depends on context length and this model’s published architecture.
| Weights (Q4_K_M) | 18.8 GB |
| Runtime overhead | 1.7 GB |
| Estimated total | 20.5 GB |
| Usable VRAM (95% of 24 GB) | 22.8 GB |
Across every GPU we track
- RTX 3060 12GBNot enough-9.1 GB
- RTX 4060 Ti 16GBNot enough-5.3 GB
- RTX 4070 Ti SUPERNot enough-5.3 GB
- RTX 3090Recommended+2.3 GB
- RTX 4090Recommended+2.3 GB
- RTX 5090Recommended+9.9 GB
- A100 40GBExcellent+17.5 GB
- RTX A6000Excellent+25.1 GB
- L40SExcellent+25.1 GB
- A100 80GBExcellent+55.5 GB
- H100 80GBExcellent+55.5 GB
- H200 141GBExcellent+113.5 GB
Estimates, not guarantees: real usage moves with runtime, driver, batch size and context. Size a specific context window in the GPU Memory Calculator.
How was this number calculated? Every figure here comes from Bitpute’s documented calculation methodology — parameters, precision, quantization, runtime overhead and usable VRAM, each formula written out in full.
Memory by quantization
| Quant | Weights | + runtime overhead |
|---|---|---|
| FP16 | 62.0 GB | 65.9 GB |
| Q8_0 | 33.0 GB | 35.3 GB |
| Q6_K | 25.4 GB | 27.5 GB |
| Q5_K_M | 22.1 GB | 23.9 GB |
| Q4_K_M | 18.8 GB | 20.5 GB |
| Q4_0 | 17.4 GB | 19.1 GB |
Overhead = 0.75 GB + 5% of weights (CUDA context, buffers). Add KV cache on top. A full 8K window is a real share of the budget.
Which quantization should you actually run?
For WizardCoder 33B v1.1, the single-GPU entry point is Q4_K_M (20.5 GB) on RTX 4090, though context headroom gets tight. Higher quality (Q6_K 27.5 GB, FP16 65.9 GB) needs RTX A6000 or multiple GPUs — for occasional use, renting is usually cheaper than buying. For a coding assistant, prefer Q5_K_M or higher: aggressive quantization tends to show up as subtle syntax and logic slips.
Single-GPU fit at Q4_K_M (20.5 GB + KV)
RTX 3060 12GBRTX 4060 Ti 16GBRTX 4070 Ti SUPERRTX 3090RTX 4090RTX 5090A100 40GBRTX A6000L40SA100 80GBH100 80GBH200 141GB
Where the memory goes
Weights and overhead are exact for Q4_K_M. The KV bar assumes a generic 32-layer transformer at 8K context — your model’s layer count and attention scheme move it, which is what the calculator is for.
Weighing the cheapest card that fits against the next one up? RTX 3090 vs RTX 5090 compares them on bandwidth, power and which models each one holds.
Common questions
How much VRAM does WizardCoder 33B v1.1 need?
WizardCoder 33B v1.1 needs about 20.5 GB of VRAM at Q4_K_M, 35.3 GB at Q8_0, or 65.9 GB at FP16. That is weights plus about 0.75 GB of runtime overhead and 5% of weight size; KV cache is additional and depends on context length.
What GPU can run WizardCoder 33B v1.1?
At Q4_K_M the smallest card in our set that fits is the RTX 3090 (24 GB usable at 95%). At FP16 you need the A100 80GB (80 GB) or larger.
Can WizardCoder 33B v1.1 run on 24 GB?
Yes. At Q4_K_M it needs about 20.5 GB, which fits inside the ~22.8 GB usable on a 24 GB card.