qwen2.5:14b on A6000, A100, H100 — 2026-04-17
2026-04-17T22:14:23.000Z → 2026-04-17T22:15:48.594Z
On 2026-04-17, qwen2.5:14b ran across A6000, A100, H100. H100 finished first at 107.9 tok/s, 13× faster than A100. The per-million-token cost: $6.41 (H100) vs $25.49 (A100) — the surprise: the headline GPU H100 was also the cheapest per-million tokens, coming in 4.0× less than A100.
Podium
H100
1st
thunder-h100
- peak tok/s
- 107.9
- avg tok/s
- 107.9
- $ / 1M tok
- $6.41
A6000
2nd
thunder-a6000
- peak tok/s
- 43.0
- avg tok/s
- 43.0
- $ / 1M tok
- $2.26
A100
3rd
thunder-a100
- peak tok/s
- 8.5
- avg tok/s
- 8.5
- $ / 1M tok
- $25.49