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

llama3.2:1b on A6000, A100, H100 — 2026-04-18

2026-04-18T02:25:49.000Z 2026-04-18T02:26:26.555Z
tweet →

On 2026-04-18, llama3.2:1b ran across A6000, A100, H100. H100 finished first at 587.8 tok/s, 22× faster than A6000. The per-million-token cost: $1.18 (H100) vs $3.61 (A6000) — the surprise: the headline GPU H100 was also the cheapest per-million tokens, coming in 3.1× less than A6000.

Podium

H100
1st
thunder-h100
peak tok/s
587.8
avg tok/s
587.8
$ / 1M tok
$1.18
A100
2nd
thunder-a100
peak tok/s
52.5
avg tok/s
52.5
$ / 1M tok
$4.13
A6000
3rd
thunder-a6000
peak tok/s
26.9
avg tok/s
26.9
$ / 1M tok
$3.61

pit wallH100 just posted 193 tok/s on llama3.1:8b — 13.7× faster than the A100 lane (1428h ago).

llama3.2:1b on A6000, A100, H100 — 2026-04-18 — Eidos Live · Crucible