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

Can I run Swift-Qwen3.8-27b on an NVIDIA B300?

Yes

Its published weights (55.6 GB) fit in the 259 GB an NVIDIA B300 can use, with room for ≈ 256K tokens of context.

Full analysis of Swift-Qwen3.8-27b Everything that runs on NVIDIA B300 Get the weights on Hugging Face
hf download ukisai/Swift-Qwen3.8-27b

Files and context on NVIDIA B300

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 NVIDIA B300 Context (16-bit)Context (8-bit)
As published55.6 GBFits256K256K
F16ukisai/Swift-Qwen3.8-27B-GGUF54.7 GBFits256K256K
Q8_0ukisai/Swift-Qwen3.8-27B-GGUF29.1 GBFits256K256K
Q6_K_Lukisai/Swift-Qwen3.8-27B-GGUF25.2 GBFits256K256K
Q6_K_Sukisai/Swift-Qwen3.8-27B-GGUF23.1 GBFits256K256K
Q6_Kukisai/Swift-Qwen3.8-27B-GGUF22.9 GBFits256K256K
Q5_K_Mukisai/Swift-Qwen3.8-27B-GGUF20.2 GBFits256K256K
Q5_K_Sukisai/Swift-Qwen3.8-27B-GGUF19.8 GBFits256K256K
Q4_K_Lukisai/Swift-Qwen3.8-27B-GGUF19.0 GBFits256K256K
Q4_K_Mukisai/Swift-Qwen3.8-27B-GGUF18.0 GBFits256K256K
IQ4_NLukisai/Swift-Qwen3.8-27B-GGUF17.7 GBFits256K256K
Q4_1ukisai/Swift-Qwen3.8-27B-GGUF17.5 GBFits256K256K
Q4_K_Sukisai/Swift-Qwen3.8-27B-GGUF16.6 GBFits256K256K
Q4_0ukisai/Swift-Qwen3.8-27B-GGUF16.0 GBFits256K256K
IQ4_XSukisai/Swift-Qwen3.8-27B-GGUF15.7 GBFits256K256K
IQ3_Mukisai/Swift-Qwen3.8-27B-GGUF15.1 GBFits256K256K
Q3_K_Lukisai/Swift-Qwen3.8-27B-GGUF14.3 GBFits256K256K
Q3_K_Mukisai/Swift-Qwen3.8-27B-GGUF13.6 GBFits256K256K
IQ3_XSukisai/Swift-Qwen3.8-27B-GGUF13.0 GBFits256K256K
Q3_K_Sukisai/Swift-Qwen3.8-27B-GGUF13.0 GBFits256K256K
IQ3_XXSukisai/Swift-Qwen3.8-27B-GGUF12.5 GBFits256K256K
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 GBFits256K256K
Q8_0bartowski/ukisai_Swift-Qwen3.8-27b-GGUF29.1 GBFits256K256K
Q6_K_Lbartowski/ukisai_Swift-Qwen3.8-27b-GGUF25.0 GBFits256K256K
Q6_Kbartowski/ukisai_Swift-Qwen3.8-27b-GGUF23.9 GBFits256K256K
Q6_K_Sbartowski/ukisai_Swift-Qwen3.8-27b-GGUF22.9 GBFits256K256K
Q5_K_Mbartowski/ukisai_Swift-Qwen3.8-27b-GGUF20.9 GBFits256K256K
Q5_K_Sbartowski/ukisai_Swift-Qwen3.8-27b-GGUF19.6 GBFits256K256K
Q4_K_Lbartowski/ukisai_Swift-Qwen3.8-27b-GGUF18.8 GBFits256K256K
Q4_1bartowski/ukisai_Swift-Qwen3.8-27b-GGUF17.8 GBFits256K256K
IQ4_NLbartowski/ukisai_Swift-Qwen3.8-27b-GGUF17.4 GBFits256K256K
Q4_K_Mbartowski/ukisai_Swift-Qwen3.8-27b-GGUF17.4 GBFits256K256K
Q4_K_Sbartowski/ukisai_Swift-Qwen3.8-27b-GGUF16.4 GBFits256K256K
Q4_0bartowski/ukisai_Swift-Qwen3.8-27b-GGUF16.3 GBFits256K256K
IQ4_XSbartowski/ukisai_Swift-Qwen3.8-27b-GGUF15.5 GBFits256K256K
IQ3_Mbartowski/ukisai_Swift-Qwen3.8-27b-GGUF14.9 GBFits256K256K
Q3_K_Lbartowski/ukisai_Swift-Qwen3.8-27b-GGUF14.1 GBFits256K256K
Q3_K_Mbartowski/ukisai_Swift-Qwen3.8-27b-GGUF13.4 GBFits256K256K
IQ3_XSbartowski/ukisai_Swift-Qwen3.8-27b-GGUF12.8 GBFits256K256K
Q3_K_Sbartowski/ukisai_Swift-Qwen3.8-27b-GGUF12.7 GBFits256K256K
IQ3_XXSbartowski/ukisai_Swift-Qwen3.8-27b-GGUF12.3 GBFits256K256K
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 NVIDIA B300

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.