llama3.1:8b on A6000, A100, H100 — 2026-04-18
2026-04-18T04:20:49.000Z → 2026-04-18T04:21:49.022Z
On 2026-04-18, llama3.1:8b ran across A6000, A100, H100. H100 finished first at 192.0 tok/s, 26× faster than A6000. The per-million-token cost: $3.60 (H100) vs $13.32 (A6000) — the surprise: the headline GPU H100 was also the cheapest per-million tokens, coming in 3.7× less than A6000.
Podium
H100
1st
thunder-h100
- peak tok/s
- 192.0
- avg tok/s
- 192.0
- $ / 1M tok
- $3.60
A100
2nd
thunder-a100
- peak tok/s
- 104.6
- avg tok/s
- 104.6
- $ / 1M tok
- $2.07
A6000
3rd
thunder-a6000
- peak tok/s
- 7.3
- avg tok/s
- 7.3
- $ / 1M tok
- $13.32