Can GLM 4.6V Flash run on the RTX 3080 10GB?
Yes. At 8K context the best fit is Q5_K_M, using about 7.6 GB of the 9.5 GB usable on the RTX 3080 10GB. Estimated speed ceiling: ~108 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | 128K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | 4.6 GB Fits | 4.7 GB Fits | 5.7 GB Fits | 9.4 GB Tight | 190 |
| 3-bit (Q3_K_M) • | 5.5 GB Fits | 5.6 GB Fits | 6.6 GB Fits | 10.3 GB No | 153 |
| 4-bit (Q4_K_M) • | 6.6 GB Fits | 6.7 GB Fits | 7.7 GB Fits | 11.4 GB No | 123 |
| 5-bit (Q5_K_M) • | 7.4 GB Fits | 7.6 GB Fits | 8.5 GB Fits | 12.3 GB No | 108 |
| 6-bit (Q6_K) • | 8.5 GB Fits | 8.7 GB Tight | 9.6 GB No | 13.4 GB No | 92 |
| 8-bit (Q8_0) • | 10.2 GB No | 10.3 GB No | 11.3 GB No | 15.0 GB No | — |
| 16-bit (F16/BF16) • | 18.4 GB No | 18.5 GB No | 19.5 GB No | 23.2 GB No | — |
• real GGUF file size from unsloth/GLM-4.6V-Flash-GGUF. Speed is a bandwidth ceiling (760 GB/s), not a benchmark. Methodology.
More
GLM 4.6V Flash on every GPU and Mac · Every model that runs on the RTX 3080 10GB · Other models that fit 12 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (zai-org/GLM-4.6V-Flash), NVIDIA specs. Data as of 2026-11-17.