Can the RTX 4090 run OTel-2.0-LLM-31B-IT-QLoRA?
Yes. OTel-2.0-LLM-31B-IT-QLoRA fits the RTX 4090's 24 GB at Q4_K_M (about 19.8 GB needed), running at an estimated 35 tok/s. There's room for up to Q5_K_M if you want higher quality.
Every quantization
OTel-2.0-LLM-31B-IT-QLoRA (31.3B) on 24 GB
| Quant | Weights | Needs | Fit | Est. speed |
|---|---|---|---|---|
| Q4_K_M | 18 GB | 19.8 GB | Fits | ~35 tok/s |
| Q5_K_M | 22 GB | 23.8 GB | Tight | ~28 tok/s |
| Q8_0 | 33 GB | 34.8 GB | Too big | - |
| FP16 | 63 GB | 64.8 GB | Too big | - |
Estimates. Needs = weights + 1.8 GB for KV cache and runtime at modest context. Speed is single-stream decode scaled from the RTX 4090's 1008 GB/s bandwidth (XiongjieDai/GPU-Benchmarks-on-LLM-Inference). Long contexts need more memory and run slower.
Compare
Same model, other hardware - and other models on this card
OTel-2.0-LLM-31B-IT-QLoRA on other hardware
- RTX 5090~57 tok/s
- RTX 5090 D~57 tok/s
- RTX 6000D~50 tok/s
- Radeon RX 7900 XTX~31 tok/s
- RTX 3090~30 tok/s
- MacBook Pro 16" M4 Max~17 tok/s
- Arc Pro B60~15 tok/s
- DGX Spark (GB10)~8.7 tok/s
Also runs on the RTX 4090
- Qwen3.6 27B~39 tok/s
- Qwen3.8 27B~39 tok/s
- Qwen3.8-27B-OBLITERATED~39 tok/s
- Gemma 4 26B A4B~184 tok/s
- gpt-oss-20b~208 tok/s
- Muse Glimmer~63 tok/s
- Nemotron 3.5 Lightning~63 tok/s
- Qwen3 14B~78 tok/s
Frequently asked
Quick answers
Can the NVIDIA RTX 4090 run OTel-2.0-LLM-31B-IT-QLoRA?
Yes. OTel-2.0-LLM-31B-IT-QLoRA fits the RTX 4090's 24 GB at Q4_K_M (about 19.8 GB needed), running at an estimated 35 tok/s. There's room for up to Q5_K_M if you want higher quality.
How much VRAM does OTel-2.0-LLM-31B-IT-QLoRA need?
About 18 GB of weights at Q4_K_M, 33 GB at Q8_0 and 63 GB at FP16. Add roughly 1.8 GB for the KV cache and runtime at modest context lengths.
How fast is OTel-2.0-LLM-31B-IT-QLoRA on the RTX 4090?
Roughly 35 tokens per second at Q4_K_M (single user, llama.cpp-style decode). This is an estimate from the RTX 4090's 1008 GB/s memory bandwidth.