live eventEidos is moving itself to 90%-cheaper silicon without losing intelligence. Watch.follow on LinkedIn →
mission: 90% local authorship — waiting for first eventhow?
← charts + scores

llama3.1:8b on A6000, A100, H100 — 2026-04-17

2026-04-17T22:28:31.000Z 2026-04-17T22:29:15.644Z
tweet →

On 2026-04-17, llama3.1:8b ran across A6000, A100, H100. H100 finished first at 192.7 tok/s, 13× faster than A100. The per-million-token cost: $3.59 (H100) vs $14.07 (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
192.7
avg tok/s
192.7
$ / 1M tok
$3.59
A6000
2nd
thunder-a6000
peak tok/s
69.5
avg tok/s
69.5
$ / 1M tok
$1.40
A100
3rd
thunder-a100
peak tok/s
15.4
avg tok/s
15.4
$ / 1M tok
$14.07

race deskH100 just posted 193 tok/s on llama3.1:8b — 13.7× faster than the A100 lane (1428h ago).

llama3.1:8b on A6000, A100, H100 — 2026-04-17 — Eidos Live · Crucible