The RTX 6000 Ada 48GB is a NVIDIA accelerator with 48 GB GDDR6 and about 960 GB/s peak memory bandwidth. AI Hashrate estimates local LLM decode (batch 1) at 8K context: fit counts weights + KV cache + 1 GB runtime overhead — not weights alone. At Q4, about 39 curated models fit fully on this card; at FP16, about 25 fit. Top Q4 speeds: Qwen3-0.6B ≈ 763.6 tok/s (estimated, fits); gemma-3-1b-it ≈ 610.9 tok/s (estimated, fits); Llama-3.2-1B-Instruct ≈ 509.1 tok/s (estimated, fits); Qwen3-1.7B ≈ 305.5 tok/s (estimated, fits); Qwen3-30B-A3B ≈ 203.6 tok/s (estimated, fits). Relative ranking is more reliable than absolute tok/s. See methodology for the bandwidth formula and measured-anchor policy. Methodology.
| Model | Params (B) | Quant | tok/s | Fits? |
|---|---|---|---|---|
| Qwen3-0.6B | 0.8 | Q4 | 763.6 est. | Yes |
| Qwen3-Next-80B-A3B-Instruct | 81.3 | Q4 | 624.8 est. | No |
| gemma-3-1b-it | 1.0 | Q4 | 610.9 est. | Yes |
| Llama-3.2-1B-Instruct | 1.2 | Q4 | 509.1 est. | Yes |
| Qwen3-1.7B | 2.0 | Q4 | 305.5 est. | Yes |
| Qwen3-0.6B | 0.8 | FP16 | 210.0 est. | Yes |
| Qwen3-30B-A3B | 30.5 | Q4 | 203.6 est. | Yes |
| Qwen3-Coder-30B-A3B-Instruct | 30.5 | Q4 | 203.6 est. | Yes |
| Qwen3-30B-A3B-Instruct-2507 | 30.5 | Q4 | 203.6 est. | Yes |
| Qwen3-Coder-Next | 79.7 | Q4 | 203.6 est. | Yes |
| Llama-3.2-3B-Instruct | 3.2 | Q4 | 190.9 est. | Yes |
| gpt-oss-20b | 21.5 | Q4 | 169.7 est. | Yes |
| gemma-3-1b-it | 1.0 | FP16 | 168.0 est. | Yes |
| Mistral-7B-Instruct-v0.3 | 7.2 | Q4 | 166.6 measured (storagereview-rtx5080-review) | Yes |
| Phi-4-mini-instruct | 3.8 | Q4 | 160.8 est. | Yes |
| Phi-4-mini-reasoning | 3.8 | Q4 | 160.8 est. | Yes |
| gemma-3-4b-it | 4.3 | Q4 | 156.6 est. | Yes |
| Qwen3-4B-Instruct-2507 | 4.0 | Q4 | 152.7 est. | Yes |
| Qwen3-4B-Thinking-2507 | 4.0 | Q4 | 152.7 est. | Yes |
| Llama-3.2-1B-Instruct | 1.2 | FP16 | 140.0 est. | Yes |
| Llama-3.1-8B-Instruct | 8.0 | Q4 | 138.6 measured (storagereview-rtx5080-review) | Yes |
| gpt-oss-120b | 120.4 | Q4 | 106.1 est. | No |
| Qwen3-30B-A3B | 30.5 | FP16 | 94.1 est. | No |
| Qwen3-Coder-30B-A3B-Instruct | 30.5 | FP16 | 94.1 est. | No |
| Qwen3-30B-A3B-Instruct-2507 | 30.5 | FP16 | 94.1 est. | No |
| Qwen3-1.7B | 2.0 | FP16 | 84.0 est. | Yes |
| GLM-4.5-Air | 110.5 | Q4 | 81.4 est. | No |
| Qwen2.5-7B-Instruct | 7.6 | Q4 | 80.4 est. | Yes |
| Qwen2.5-Coder-7B-Instruct | 7.6 | Q4 | 80.4 est. | Yes |
| DeepSeek-R1-Distill-Qwen-7B | 7.6 | Q4 | 80.4 est. | Yes |
| Qwen3-8B | 8.2 | Q4 | 74.5 est. | Yes |
| DeepSeek-R1-0528-Qwen3-8B | 8.2 | Q4 | 74.5 est. | Yes |
| Llama-3.2-3B-Instruct | 3.2 | FP16 | 52.5 est. | Yes |
| gemma-3-12b-it | 12.2 | Q4 | 50.9 est. | Yes |
| GLM-4.7-Flash | 31.2 | Q4 | 50.9 est. | Yes |
| Mixtral-8x7B-Instruct-v0.1 | 46.7 | Q4 | 47.4 est. | Yes |
| gpt-oss-20b | 21.5 | FP16 | 46.7 est. | Yes |
| Phi-4-mini-instruct | 3.8 | FP16 | 44.2 est. | Yes |
| Phi-4-mini-reasoning | 3.8 | FP16 | 44.2 est. | Yes |
| gemma-3-4b-it | 4.3 | FP16 | 43.1 est. | Yes |
| Qwen3-4B-Instruct-2507 | 4.0 | FP16 | 42.0 est. | Yes |
| Qwen3-4B-Thinking-2507 | 4.0 | FP16 | 42.0 est. | Yes |
| Phi-4 | 14.7 | Q4 | 41.6 est. | Yes |
| Phi-4-reasoning | 14.7 | Q4 | 41.6 est. | Yes |
| Qwen3-14B | 14.8 | Q4 | 41.3 est. | Yes |
| Qwen2.5-14B-Instruct | 14.8 | Q4 | 41.3 est. | Yes |
| Qwen2.5-Coder-14B-Instruct | 14.8 | Q4 | 41.3 est. | Yes |
| DeepSeek-R1-Distill-Qwen-14B | 14.8 | Q4 | 41.3 est. | Yes |
| DeepSeek-V4-Flash | 158.1 | Q4 | 37.7 est. | No |
| Devstral-Small-2507 | 23.6 | Q4 | 25.9 est. | Yes |
| Mistral-Small-3.2-24B-Instruct-2506 | 24.0 | Q4 | 25.5 est. | Yes |
| Magistral-Small-2509 | 24.0 | Q4 | 25.5 est. | Yes |
| MiniMax-M2.7 | 228.7 | Q4 | 24.2 est. | No |
| Mistral-7B-Instruct-v0.3 | 7.2 | FP16 | 23.3 est. | Yes |
| gemma-3-27b-it | 27.4 | Q4 | 22.5 est. | Yes |
| Qwen2.5-7B-Instruct | 7.6 | FP16 | 22.1 est. | Yes |
| Qwen2.5-Coder-7B-Instruct | 7.6 | FP16 | 22.1 est. | Yes |
| DeepSeek-R1-Distill-Qwen-7B | 7.6 | FP16 | 22.1 est. | Yes |
| GLM-4.7-Flash | 31.2 | FP16 | 21.3 est. | No |
| Llama-3.1-8B-Instruct | 8.0 | FP16 | 21.0 est. | Yes |
| Qwen3-8B | 8.2 | FP16 | 20.5 est. | Yes |
| DeepSeek-R1-0528-Qwen3-8B | 8.2 | FP16 | 20.5 est. | Yes |
| Llama-3.1-70B-Instruct | 70.6 | Q4 | 19.8 est. | No |
| Llama-3.3-70B-Instruct | 70.6 | Q4 | 19.8 est. | No |
| DeepSeek-R1-Distill-Llama-70B | 70.6 | Q4 | 19.6 est. | No |
| Qwen3-32B | 32.8 | Q4 | 18.6 est. | Yes |
| Qwen2.5-32B-Instruct | 32.8 | Q4 | 18.6 est. | Yes |
| Qwen2.5-Coder-32B-Instruct | 32.8 | Q4 | 18.6 est. | Yes |
| DeepSeek-R1-Distill-Qwen-32B | 32.8 | Q4 | 18.6 est. | Yes |
| Qwen2.5-72B-Instruct | 72.7 | Q4 | 17.9 est. | No |
| Devstral-Small-2507 | 23.6 | FP16 | 16.7 est. | No |
| Mistral-Small-3.2-24B-Instruct-2506 | 24.0 | FP16 | 15.9 est. | No |
| Magistral-Small-2509 | 24.0 | FP16 | 15.9 est. | No |
| Qwen3-Coder-Next | 79.7 | FP16 | 14.2 est. | No |
| gemma-3-12b-it | 12.2 | FP16 | 14.0 est. | Yes |
| Qwen3-Next-80B-A3B-Instruct | 81.3 | FP16 | 13.7 est. | No |
| Qwen3-14B | 14.8 | FP16 | 11.4 est. | Yes |
| Qwen2.5-14B-Instruct | 14.8 | FP16 | 11.4 est. | Yes |
| Qwen2.5-Coder-14B-Instruct | 14.8 | FP16 | 11.4 est. | Yes |
| DeepSeek-R1-Distill-Qwen-14B | 14.8 | FP16 | 11.4 est. | Yes |
| Phi-4 | 14.7 | FP16 | 11.4 est. | Yes |
| Phi-4-reasoning | 14.7 | FP16 | 11.4 est. | Yes |
| gemma-3-27b-it | 27.4 | FP16 | 10.8 est. | No |
| Qwen3-235B-A22B | 235.1 | Q4 | 10.1 est. | No |
| Qwen3-235B-A22B-Instruct-2507 | 235.1 | Q4 | 10.1 est. | No |
| Mixtral-8x7B-Instruct-v0.1 | 46.7 | FP16 | 9.1 est. | No |
| Qwen3-32B | 32.8 | FP16 | 6.3 est. | No |
| Qwen2.5-32B-Instruct | 32.8 | FP16 | 6.3 est. | No |
| Qwen2.5-Coder-32B-Instruct | 32.8 | FP16 | 6.3 est. | No |
| DeepSeek-R1-Distill-Qwen-32B | 32.8 | FP16 | 6.3 est. | No |
| Qwen2.5-72B-Instruct | 72.7 | FP16 | 0.6 est. | No |
| Llama-3.1-70B-Instruct | 70.6 | FP16 | 0.6 est. | No |
| Llama-3.3-70B-Instruct | 70.6 | FP16 | 0.6 est. | No |
| DeepSeek-R1-Distill-Llama-70B | 70.6 | FP16 | 0.6 est. | No |