The H200 141GB SXM5 is a NVIDIA accelerator with 141 GB HBM3e and about 4800 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 48 curated models fit fully on this card; at FP16, about 38 fit. Top Q4 speeds: Qwen3-0.6B ≈ 3818.2 tok/s (estimated, fits); gemma-3-1b-it ≈ 3054.5 tok/s (estimated, fits); Llama-3.2-1B-Instruct ≈ 2545.5 tok/s (estimated, fits); Qwen3-1.7B ≈ 1527.3 tok/s (estimated, fits); Qwen3-30B-A3B ≈ 1018.2 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 | 3818.2 est. | Yes |
| gemma-3-1b-it | 1.0 | Q4 | 3054.5 est. | Yes |
| Llama-3.2-1B-Instruct | 1.2 | Q4 | 2545.5 est. | Yes |
| Qwen3-1.7B | 2.0 | Q4 | 1527.3 est. | Yes |
| Qwen3-0.6B | 0.8 | FP16 | 1050.0 est. | Yes |
| Qwen3-30B-A3B | 30.5 | Q4 | 1018.2 est. | Yes |
| Qwen3-Coder-30B-A3B-Instruct | 30.5 | Q4 | 1018.2 est. | Yes |
| Qwen3-30B-A3B-Instruct-2507 | 30.5 | Q4 | 1018.2 est. | Yes |
| Qwen3-Next-80B-A3B-Instruct | 81.3 | Q4 | 1018.2 est. | Yes |
| Qwen3-Coder-Next | 79.7 | Q4 | 1018.2 est. | Yes |
| Llama-3.2-3B-Instruct | 3.2 | Q4 | 954.5 est. | Yes |
| gpt-oss-20b | 21.5 | Q4 | 848.5 est. | Yes |
| gemma-3-1b-it | 1.0 | FP16 | 840.0 est. | Yes |
| Phi-4-mini-instruct | 3.8 | Q4 | 803.8 est. | Yes |
| Phi-4-mini-reasoning | 3.8 | Q4 | 803.8 est. | Yes |
| gemma-3-4b-it | 4.3 | Q4 | 783.2 est. | Yes |
| Qwen3-4B-Instruct-2507 | 4.0 | Q4 | 763.6 est. | Yes |
| Qwen3-4B-Thinking-2507 | 4.0 | Q4 | 763.6 est. | Yes |
| Llama-3.2-1B-Instruct | 1.2 | FP16 | 700.0 est. | Yes |
| Qwen3-Coder-Next | 79.7 | FP16 | 613.6 est. | No |
| Qwen3-Next-80B-A3B-Instruct | 81.3 | FP16 | 589.9 est. | No |
| Qwen3-235B-A22B | 235.1 | Q4 | 434.6 est. | No |
| Qwen3-235B-A22B-Instruct-2507 | 235.1 | Q4 | 434.6 est. | No |
| Mistral-7B-Instruct-v0.3 | 7.2 | Q4 | 424.2 est. | Yes |
| Qwen3-1.7B | 2.0 | FP16 | 420.0 est. | Yes |
| Qwen2.5-7B-Instruct | 7.6 | Q4 | 401.9 est. | Yes |
| Qwen2.5-Coder-7B-Instruct | 7.6 | Q4 | 401.9 est. | Yes |
| DeepSeek-R1-Distill-Qwen-7B | 7.6 | Q4 | 401.9 est. | Yes |
| Llama-3.1-8B-Instruct | 8.0 | Q4 | 381.8 est. | Yes |
| gpt-oss-120b | 120.4 | Q4 | 381.8 est. | Yes |
| Qwen3-8B | 8.2 | Q4 | 372.5 est. | Yes |
| DeepSeek-R1-0528-Qwen3-8B | 8.2 | Q4 | 372.5 est. | Yes |
| MiniMax-M2.7 | 228.7 | Q4 | 305.5 est. | Yes |
| Qwen3-30B-A3B | 30.5 | FP16 | 280.0 est. | Yes |
| Qwen3-Coder-30B-A3B-Instruct | 30.5 | FP16 | 280.0 est. | Yes |
| Qwen3-30B-A3B-Instruct-2507 | 30.5 | FP16 | 280.0 est. | Yes |
| Llama-3.2-3B-Instruct | 3.2 | FP16 | 262.5 est. | Yes |
| gemma-3-12b-it | 12.2 | Q4 | 254.5 est. | Yes |
| GLM-4.7-Flash | 31.2 | Q4 | 254.5 est. | Yes |
| GLM-4.5-Air | 110.5 | Q4 | 254.5 est. | Yes |
| Mixtral-8x7B-Instruct-v0.1 | 46.7 | Q4 | 236.8 est. | Yes |
| DeepSeek-V4-Flash | 158.1 | Q4 | 235.0 est. | Yes |
| gpt-oss-20b | 21.5 | FP16 | 233.3 est. | Yes |
| Phi-4-mini-instruct | 3.8 | FP16 | 221.1 est. | Yes |
| Phi-4-mini-reasoning | 3.8 | FP16 | 221.1 est. | Yes |
| gemma-3-4b-it | 4.3 | FP16 | 215.4 est. | Yes |
| Qwen3-4B-Instruct-2507 | 4.0 | FP16 | 210.0 est. | Yes |
| Qwen3-4B-Thinking-2507 | 4.0 | FP16 | 210.0 est. | Yes |
| Phi-4 | 14.7 | Q4 | 207.8 est. | Yes |
| Phi-4-reasoning | 14.7 | Q4 | 207.8 est. | Yes |
| Qwen3-14B | 14.8 | Q4 | 206.4 est. | Yes |
| Qwen2.5-14B-Instruct | 14.8 | Q4 | 206.4 est. | Yes |
| Qwen2.5-Coder-14B-Instruct | 14.8 | Q4 | 206.4 est. | Yes |
| DeepSeek-R1-Distill-Qwen-14B | 14.8 | Q4 | 206.4 est. | Yes |
| GLM-4.5 | 358.3 | Q4 | 129.7 est. | No |
| Devstral-Small-2507 | 23.6 | Q4 | 129.4 est. | Yes |
| Mistral-Small-3.2-24B-Instruct-2506 | 24.0 | Q4 | 127.3 est. | Yes |
| Magistral-Small-2509 | 24.0 | Q4 | 127.3 est. | Yes |
| Mistral-7B-Instruct-v0.3 | 7.2 | FP16 | 116.7 est. | Yes |
| gemma-3-27b-it | 27.4 | Q4 | 112.3 est. | Yes |
| Qwen2.5-7B-Instruct | 7.6 | FP16 | 110.5 est. | Yes |
| Qwen2.5-Coder-7B-Instruct | 7.6 | FP16 | 110.5 est. | Yes |
| DeepSeek-R1-Distill-Qwen-7B | 7.6 | FP16 | 110.5 est. | Yes |
| Llama-3.1-8B-Instruct | 8.0 | FP16 | 105.0 est. | Yes |
| Qwen3-8B | 8.2 | FP16 | 102.4 est. | Yes |
| DeepSeek-R1-0528-Qwen3-8B | 8.2 | FP16 | 102.4 est. | Yes |
| gpt-oss-120b | 120.4 | FP16 | 100.7 est. | No |
| Qwen3-32B | 32.8 | Q4 | 93.1 est. | Yes |
| Qwen2.5-32B-Instruct | 32.8 | Q4 | 93.1 est. | Yes |
| Qwen2.5-Coder-32B-Instruct | 32.8 | Q4 | 93.1 est. | Yes |
| DeepSeek-R1-Distill-Qwen-32B | 32.8 | Q4 | 93.1 est. | Yes |
| GLM-4.5-Air | 110.5 | FP16 | 79.0 est. | No |
| gemma-3-12b-it | 12.2 | FP16 | 70.0 est. | Yes |
| GLM-4.7-Flash | 31.2 | FP16 | 70.0 est. | Yes |
| Qwen3-Coder-480B-A35B-Instruct | 480.2 | Q4 | 67.0 est. | No |
| Mixtral-8x7B-Instruct-v0.1 | 46.7 | FP16 | 65.1 est. | Yes |
| Phi-4 | 14.7 | FP16 | 57.1 est. | Yes |
| Phi-4-reasoning | 14.7 | FP16 | 57.1 est. | Yes |
| Qwen3-14B | 14.8 | FP16 | 56.8 est. | Yes |
| Qwen2.5-14B-Instruct | 14.8 | FP16 | 56.8 est. | Yes |
| Qwen2.5-Coder-14B-Instruct | 14.8 | FP16 | 56.8 est. | Yes |
| DeepSeek-R1-Distill-Qwen-14B | 14.8 | FP16 | 56.8 est. | Yes |
| MiniMax-M1 | 456.1 | Q4 | 55.5 est. | No |
| Llama-3.1-70B-Instruct | 70.6 | Q4 | 43.6 est. | Yes |
| Llama-3.3-70B-Instruct | 70.6 | Q4 | 43.6 est. | Yes |
| DeepSeek-R1-Distill-Llama-70B | 70.6 | Q4 | 43.3 est. | Yes |
| Qwen2.5-72B-Instruct | 72.7 | Q4 | 42.0 est. | Yes |
| DeepSeek-V4-Flash | 158.1 | FP16 | 35.9 est. | No |
| Devstral-Small-2507 | 23.6 | FP16 | 35.6 est. | Yes |
| Mistral-Small-3.2-24B-Instruct-2506 | 24.0 | FP16 | 35.0 est. | Yes |
| Magistral-Small-2509 | 24.0 | FP16 | 35.0 est. | Yes |
| DeepSeek-R1 | 684.5 | Q4 | 31.7 est. | No |
| DeepSeek-V3.2 | 685.4 | Q4 | 31.6 est. | No |
| gemma-3-27b-it | 27.4 | FP16 | 30.9 est. | Yes |
| Llama-3.1-70B-Instruct | 70.6 | FP16 | 27.9 est. | No |
| Llama-3.3-70B-Instruct | 70.6 | FP16 | 27.9 est. | No |
| DeepSeek-R1-Distill-Llama-70B | 70.6 | FP16 | 27.7 est. | No |
| Qwen3-32B | 32.8 | FP16 | 25.6 est. | Yes |
| Qwen2.5-32B-Instruct | 32.8 | FP16 | 25.6 est. | Yes |
| Qwen2.5-Coder-32B-Instruct | 32.8 | FP16 | 25.6 est. | Yes |
| DeepSeek-R1-Distill-Qwen-32B | 32.8 | FP16 | 25.6 est. | Yes |
| Qwen2.5-72B-Instruct | 72.7 | FP16 | 25.3 est. | No |
| GLM-5.1 | 753.9 | Q4 | 24.2 est. | No |
| GLM-5.2 | 753.3 | Q4 | 24.2 est. | No |
| MiniMax-M2.7 | 228.7 | FP16 | 22.5 est. | No |
| DeepSeek-V4-Pro | 861.6 | Q4 | 15.1 est. | No |
| Qwen3-235B-A22B | 235.1 | FP16 | 9.6 est. | No |
| Qwen3-235B-A22B-Instruct-2507 | 235.1 | FP16 | 9.6 est. | No |
| Llama-3.1-405B-Instruct | 405.9 | Q4 | 4.6 est. | No |