llama3.1:8b on A100, H100 — 2026-04-18
2026-04-18T06:28:57.000Z → 2026-04-18T06:29:17.981Z
On 2026-04-18, llama3.1:8b ran across A100, H100. H100 finished first at 192.2 tok/s, 12× faster than A100. The per-million-token cost: $3.60 (H100) vs $13.89 (A100) — the surprise: the headline GPU H100 was also the cheapest per-million tokens, coming in 3.9× less than A100.
Podium
H100
1st
thunder-h100
- peak tok/s
- 192.2
- avg tok/s
- 192.2
- $ / 1M tok
- $3.60
A100
2nd
thunder-a100
- peak tok/s
- 15.6
- avg tok/s
- 15.6
- $ / 1M tok
- $13.89