Every 7 days the grams on this network work the model catalog in cohorts — each family's lead member at rung 0 on every engine first, then its quantization ladder, then the other members, then derivatives — and publish what they proved, per hardware class, with how many machines contributed. No number here is from a model card. Publishers are listed by what the network proved of their builds, never by popularity.
2026-W46
Open now. Grams take the least-covered artifact for their own hardware class first, so two machines of one kind never measure the same thing while a third kind waits.
code.reason, code.write, embed.vector, faithful_tool_result, instruct.follow, lang.as, lang.bn, lang.en, lang.es, lang.hi, lang.ja, lang.kn, lang.ne, lang.or …
vision.count, vision.spatial
voice.identify
audio, audio.in, audio.out, lang.as, lang.bn, lang.en, lang.gu, lang.kn, lang.ml, lang.mr, lang.or, lang.pa, lang.ta, lang.te
audio.music
image.generate
translate.as, translate.bn, translate.gu, translate.hi, translate.kn, translate.ml, translate.mr, translate.or, translate.pa, translate.ta, translate.te, translate.text, translate.ur
per engine: how many artifacts this season proved, on how many hardware classes, which capabilities came back, and how many artifact-slots the fleet currently cannot reach at all (no Apple Silicon for MLX, no schedulable card for vLLM, no cluster for the speech pods). An engine with zero proved and a large unreachable count has never been shown to work on this network -- a statement about the network, not about the models.
away 133 h · last session stopped on purpose — shut down, or restarted to take a build
expects 326 h for the rest (own samples) · trailing rate says 4004 h · blocked: fetch
away 298 h · last session ended without being asked to — power loss, a hard reset, or a crash
expects 427 h for the rest (own samples) · trailing rate says 7794 h · blocked: fetch
away 442 h
expects 221 h for the rest (own samples) · trailing rate says 1388 h · blocked: fetch
away 319 h · last session ended without being asked to — power loss, a hard reset, or a crash
expects 1030 h for the rest (own samples) · trailing rate says 16260 h · blocked: pace
expects 120 h for the rest (own samples) · trailing rate says 6570 h · blocked: engine
away 54 h · last session a supervisor bounce: a new session started while the last one was alive
expects 88 h for the rest (own samples) · trailing rate says 11136 h · blocked: engine
expects 1483 h for the rest (own samples) · blocked: budget
away 180 h · last session ended without being asked to — power loss, a hard reset, or a crash
expects 299 h for the rest (own samples) · blocked: pace
away 36 min · last session a supervisor bounce: a new session started while the last one was alive
expects 96 h for the rest (own samples) · blocked: pace
away 345 h · last session a supervisor bounce: a new session started while the last one was alive · a laptop on battery — most likely asleep
expects 357 h for the rest (own samples) · trailing rate says 4470 h · blocked: fetch
away 234 h
expects 286 h for the rest (machines of this class) · trailing rate says 14024 h · blocked: engine
away 246 h · last session ended without being asked to — power loss, a hard reset, or a crash · this machine sleeps (43 min during its last session) and is on battery — most likely asleep
expects 192 h for the rest (own samples) · blocked: budget
away 231 h
expects 26 h for the rest (own samples) · trailing rate says 12648 h · blocked: engine
expects 554 h for the rest (machines of this class) · blocked: budget
away 334 h
expects 0 h for the rest (machines of this class) · trailing rate says 33048 h · blocked: pace
away 582 h · last session ended without being asked to — power loss, a hard reset, or a crash
no priced estimate yet · trailing rate says 3336 h · blocked: pace
away 527 h · last session ended without being asked to — power loss, a hard reset, or a crash
expects 112 h for the rest (own samples) · blocked: pace
away 331 h
expects 0 h for the rest (machines of this class) · blocked: pace
away 415 h · last session a supervisor bounce: a new session started while the last one was alive · a laptop on battery — most likely asleep
no priced estimate yet · blocked: pace
away 319 h
expects 0 h for the rest (machines of this class) · blocked: pace
away 330 h
expects 0 h for the rest (machines of this class) · blocked: pace
away 260 h · last session a supervisor bounce: a new session started while the last one was alive · this machine sleeps (4 min during its last session) and is on battery — most likely asleep
expects 137 h for the rest (own samples) · blocked: pace
away 334 h
expects 0 h for the rest (machines of this class) · blocked: pace
away 334 h
expects 0 h for the rest (machines of this class) · blocked: pace
away 331 h
expects 0 h for the rest (machines of this class) · blocked: pace
away 330 h
expects 0 h for the rest (machines of this class) · blocked: pace
away 14 min · last session ended without being asked to — power loss, a hard reset, or a crash · this machine sleeps (362 min during its last session) — most likely asleep
no priced estimate yet · blocked: pace
away 45 min · last session ended without being asked to — power loss, a hard reset, or a crash
no priced estimate yet · blocked: disk
expects 8 h for the rest (own samples) · blocked: fetch
away 18 min · last session a supervisor bounce: a new session started while the last one was alive
no priced estimate yet · blocked: pace
expects 1437 h for the rest (own samples) · blocked: pace
away 333 h
expects 0 h for the rest (machines of this class) · blocked: pace
away 207 h · last session ended without being asked to — power loss, a hard reset, or a crash
no priced estimate yet · blocked: disk
away 319 h
expects 0 h for the rest (machines of this class) · blocked: pace
one row per gram: what it is, what its hardware can never reach, how far it has got, and its OWN priced estimate for the rest. There is no single fleet ETA, because a gram only proves what it can obtain.
proving is a LEASE, not a fact: what each gram said it was working on when it last reported. It expires on its own, confers nothing and is never coverage. It is here so the network can be watched working.
116
cells signed as gated, unloadable, or not published upstream. Each one is a measurement every other machine of that class is spared.
Medians from the grams own clocks, carried across windows so a class whose machines are asleep still plans on what the network paid to learn.
Tokens actually served, cumulative. Demand decides how deep a proof goes and where weights sit — never what exists.
Past seasons
The season files are public JSON, built by an open aggregator from signed results anyone can re-verify — data/catalog/seasons on GitHub. Cite them; they carry the signatures. Back to the evidence table →