toktape — published LLM inference runs and their caveats

toktapethe record, not a recording of it

Every run on this page is the file toktape wrote on somebody's own server: the card, the caveats that say when a figure is not quotable, and a replay, all drawn from those same bytes. One word on the machine already running llama-server records yours.

brew install midagedev/tap/toktape other ways to install GitHub

Narrowed toquant: iq3_s ×engine: bloomery ×Clear all

4 runs, newest first

bloomery · MiniMax-M3 UD-IQ4_XS · RTX A6000 + host experts · 한국어 (전세·월세) 14.5 tok/sprefill 37.7 tok/s

MiniMax-M3-UD-IQ4_XS-00001-of-00006.gguf

bloomery0.3.6 (93c017bd-dirty)iq3_sP197 · G512 · cache 0%ctx 128koffload partialrtx-a6000 linux 1 stream

midagedev

2026-10-21

GLM-5.3-Flash-UD-IQ4_XS-00001-of-00005.gguf 24.6 tok/s

bloomery0.2.10 (8e55b8ce)iq3_sP38 · G400 · cache 0%ctx 8koffload partialrtx-a6000 linux 2 streams ! 4 caveats

midagedev

2026-10-10

GLM-5.3-Flash-UD-IQ4_XS-00001-of-00005.gguf 25.1 tok/s

bloomery0.2.10 (8e55b8ce)iq3_sP33 · G400 · cache 0%ctx 16koffload partialrtx-a6000 linux 1 stream ! 3 caveats

midagedev

2026-10-10

GLM-5.3-Flash-UD-IQ4_XS-00001-of-00005.gguf 24.6 tok/sprefill 106 tok/s

bloomery0.2.10 (8e55b8ce)iq3_sP5548 · G600 · cache 0%ctx 16koffload partialrtx-a6000 linux 1 stream ! 1 caveat

midagedev

2026-10-10

Newest first unless you choose another order; there is still no rank column and no score. Every row carries the caveats the card would print, because a result set without them is a leaderboard with the sorting taken out.
toktape publish <run.toktape> puts one here. toktape on GitHub

Page views are counted with Cloudflare Web Analytics: no cookies, no fingerprinting. The toktape binary itself sends nothing anywhere.