Model Anatomy See what's inside open-weight AI models

Can I run Swift-Qwen3.8-27b on an RTX 5090?

Yes, with a GGUF version

The published weights are too big, but the Q6_K_L file from ukisai/Swift-Qwen3.8-27B-GGUF (25.2 GB) fits, with room for ≈ 52K tokens of context.

Full analysis of Swift-Qwen3.8-27b Everything that runs on RTX 5090 Get the Q6_K_L file on Hugging Face
ollama run hf.co/ukisai/Swift-Qwen3.8-27B-GGUF:Q6_K_L

Files and context on RTX 5090

Each file of Swift-Qwen3.8-27b (and of GGUF versions the site has analyzed), and how much context fits next to it, with the context cache in 16-bit or 8-bit.

FileSizeOn RTX 5090 Context (16-bit)Context (8-bit)
As published55.6 GBDoesn't fit——
F16ukisai/Swift-Qwen3.8-27B-GGUF54.7 GBDoesn't fit——
Q8_0ukisai/Swift-Qwen3.8-27B-GGUF29.1 GBDoesn't fit——
Q6_K_Lukisai/Swift-Qwen3.8-27B-GGUF25.2 GBTight52K104K
Q6_K_Sukisai/Swift-Qwen3.8-27B-GGUF23.1 GBTight83K166K
Q6_Kukisai/Swift-Qwen3.8-27B-GGUF22.9 GBFits86K172K
Q5_K_Mukisai/Swift-Qwen3.8-27B-GGUF20.2 GBFits126K252K
Q5_K_Sukisai/Swift-Qwen3.8-27B-GGUF19.8 GBFits132K256K
Q4_K_Lukisai/Swift-Qwen3.8-27B-GGUF19.0 GBFits143K256K
Q4_K_Mukisai/Swift-Qwen3.8-27B-GGUF18.0 GBFits158K256K
IQ4_NLukisai/Swift-Qwen3.8-27B-GGUF17.7 GBFits164K256K
Q4_1ukisai/Swift-Qwen3.8-27B-GGUF17.5 GBFits166K256K
Q4_K_Sukisai/Swift-Qwen3.8-27B-GGUF16.6 GBFits180K256K
Q4_0ukisai/Swift-Qwen3.8-27B-GGUF16.0 GBFits189K256K
IQ4_XSukisai/Swift-Qwen3.8-27B-GGUF15.7 GBFits193K256K
IQ3_Mukisai/Swift-Qwen3.8-27B-GGUF15.1 GBFits202K256K
Q3_K_Lukisai/Swift-Qwen3.8-27B-GGUF14.3 GBFits213K256K
Q3_K_Mukisai/Swift-Qwen3.8-27B-GGUF13.6 GBFits224K256K
IQ3_XSukisai/Swift-Qwen3.8-27B-GGUF13.0 GBFits233K256K
Q3_K_Sukisai/Swift-Qwen3.8-27B-GGUF13.0 GBFits234K256K
IQ3_XXSukisai/Swift-Qwen3.8-27B-GGUF12.5 GBFits240K256K
Q2_Kukisai/Swift-Qwen3.8-27B-GGUF11.0 GBFits256K256K
IQ2_Mukisai/Swift-Qwen3.8-27B-GGUF10.7 GBFits256K256K
IQ2_Sukisai/Swift-Qwen3.8-27B-GGUF9.90 GBFits256K256K
IQ2_XSukisai/Swift-Qwen3.8-27B-GGUF9.30 GBFits256K256K
IQ2_XXSukisai/Swift-Qwen3.8-27B-GGUF9.09 GBFits256K256K
BF16bartowski/ukisai_Swift-Qwen3.8-27b-GGUF54.7 GBDoesn't fit——
Q8_0bartowski/ukisai_Swift-Qwen3.8-27b-GGUF29.1 GBDoesn't fit——
Q6_K_Lbartowski/ukisai_Swift-Qwen3.8-27b-GGUF25.0 GBTight55K110K
Q6_Kbartowski/ukisai_Swift-Qwen3.8-27b-GGUF23.9 GBTight71K143K
Q6_K_Sbartowski/ukisai_Swift-Qwen3.8-27b-GGUF22.9 GBFits86K173K
Q5_K_Mbartowski/ukisai_Swift-Qwen3.8-27b-GGUF20.9 GBFits115K230K
Q5_K_Sbartowski/ukisai_Swift-Qwen3.8-27b-GGUF19.6 GBFits135K256K
Q4_K_Lbartowski/ukisai_Swift-Qwen3.8-27b-GGUF18.8 GBFits146K256K
Q4_1bartowski/ukisai_Swift-Qwen3.8-27b-GGUF17.8 GBFits161K256K
IQ4_NLbartowski/ukisai_Swift-Qwen3.8-27b-GGUF17.4 GBFits167K256K
Q4_K_Mbartowski/ukisai_Swift-Qwen3.8-27b-GGUF17.4 GBFits167K256K
Q4_K_Sbartowski/ukisai_Swift-Qwen3.8-27b-GGUF16.4 GBFits183K256K
Q4_0bartowski/ukisai_Swift-Qwen3.8-27b-GGUF16.3 GBFits183K256K
IQ4_XSbartowski/ukisai_Swift-Qwen3.8-27b-GGUF15.5 GBFits196K256K
IQ3_Mbartowski/ukisai_Swift-Qwen3.8-27b-GGUF14.9 GBFits205K256K
Q3_K_Lbartowski/ukisai_Swift-Qwen3.8-27b-GGUF14.1 GBFits216K256K
Q3_K_Mbartowski/ukisai_Swift-Qwen3.8-27b-GGUF13.4 GBFits227K256K
IQ3_XSbartowski/ukisai_Swift-Qwen3.8-27b-GGUF12.8 GBFits236K256K
Q3_K_Sbartowski/ukisai_Swift-Qwen3.8-27b-GGUF12.7 GBFits237K256K
IQ3_XXSbartowski/ukisai_Swift-Qwen3.8-27b-GGUF12.3 GBFits243K256K
Q2_Kbartowski/ukisai_Swift-Qwen3.8-27b-GGUF10.8 GBFits256K256K
IQ2_Mbartowski/ukisai_Swift-Qwen3.8-27b-GGUF10.5 GBFits256K256K
IQ2_Sbartowski/ukisai_Swift-Qwen3.8-27b-GGUF9.68 GBFits256K256K
IQ2_XSbartowski/ukisai_Swift-Qwen3.8-27b-GGUF9.09 GBFits256K256K
IQ2_XXSbartowski/ukisai_Swift-Qwen3.8-27b-GGUF8.88 GBFits256K256K

Swift-Qwen3.8-27b on other hardware

Popular models that run on RTX 5090

Assumes 90% of the card's memory is usable and at least 4K tokens of context for one conversation, with the context cache in 16-bit unless noted. Fits leaves at least 20% of that memory for context; Tight leaves less.