llama3.1:8b on A100, H100 — 2026-04-18
2026-04-18T07:00:55.000Z → 2026-04-18T07:01:16.271Z
On 2026-04-18, llama3.1:8b ran across A100, H100. H100 finished first at 193.2 tok/s, 14× faster than A100. The per-million-token cost: $3.58 (H100) vs $15.37 (A100) — the surprise: the headline GPU H100 was also the cheapest per-million tokens, coming in 4.3× less than A100.
Podium
H100
1st
thunder-h100
- peak tok/s
- 193.2
- avg tok/s
- 193.2
- $ / 1M tok
- $3.58
A100
2nd
thunder-a100
- peak tok/s
- 14.1
- avg tok/s
- 14.1
- $ / 1M tok
- $15.37