GPU × use-case guide · Vision-language

Is the H100 SXM a good GPU for vision-language?

The H100 SXM is a hopper NVIDIA GPU with 80 GB HBM3 (3350 GB/s), 990 TFLOPS BF16 / 1979 TFLOPS FP8, and a 700 W TDP. Vision-language workloads care most about extra VRAM for the vision encoder and image tokens on top of the language model. Here's how the H100 SXM measures up.

What models fit on a single H100 SXM?

Weights only, reserving ~25% of the 80 GB for KV cache, activations and fragmentation. ✓ = fits on one card.

Largest single-card fit: Gemma 2 27B at BF16, Mixtral 8x7B (MoE) at FP8, Qwen 2.5 72B at INT4. Bigger models need tensor-parallel across 8 cards.

H100 SXM for vision-language, specifically

Vision-language is context-heavy, so the KV cache — not the weights — is what fills the 80 GB. On the H100 SXM you'll trade context length against batch size: long prompts mean fewer concurrent requests. Because it runs offline, batch aggressively to push tokens-per-dollar down. Size it precisely on the calculator.

H100 SXM pricing across providers

Verdict

With 80 GB, the H100 SXM is a data-center-class card that comfortably handles vision-language for models up to Gemma 2 27B at full precision on a single card — a strong pick if your budget supports ~$2.49/hr.