The catalog proves itself in seasons.

• ollama20 artifacts proved on 1 class

  code.reason, code.write, embed.vector, faithful_tool_result, instruct.follow, lang.bn, lang.en, lang.gu, lang.hi, lang.kn, lang.ml, lang.mr, lang.nl, lang.pa …

• speech-hf17 artifacts proved on 1 class · 65 unreachable

  audio, audio.in, audio.out, lang.en, lang.hi, lang.mr, lang.ne, lang.ta, lang.te, lang.ur

• translate-hf4 artifacts proved on 1 class

  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

• speech-pod3 artifacts proved on 1 class · 80 unreachable

  audio, audio.in, audio.out, lang.en

• vision-task3 artifacts proved on 1 class · 18 unreachable

  vision.count, vision.spatial

• diffusers2 artifacts proved on 1 class · 60 unreachable

  image.generate

• speaker-id2 artifacts proved on 1 class · 10 unreachable

  voice.identify

• audiogennothing proved on this network yet · 12 unreachable
• mlxnothing proved on this network yet · 36 unreachable
• mlx-imagenothing proved on this network yet · 12 unreachable
• mlx-speechnothing proved on this network yet · 24 unreachable
• speech-duplexnothing proved on this network yet · 15 unreachable
• speech-hostnothing proved on this network yet · 56 unreachable
• vllmnothing proved on this network yet · 48 unreachable

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.

• ○ acbe75 · Linux · x86-64 · no GPU · 32 GiB47 proved · 1561 queued · 14 it cannot reach

  away 48 h

  expects 221 h for the rest (own samples) · trailing rate says 1388 h · blocked: fetch

• ○ e61d3b · Linux · x86-64 · no GPU · 32 GiB45 proved · 1589 queued · 14 it cannot reach

  away 2 h · last session a supervisor bounce: a new session started while the last one was alive

  expects 412 h for the rest (own samples) · trailing rate says 3814 h · blocked: fetch

• ○ a5be46 · Linux · x86-64 · NVIDIA · 32 GiB24 proved · 162 queued · 10 it cannot reach

  away 61 h · last session ended without being asked to — power loss, a hard reset, or a crash

  expects 23 h for the rest (own samples) · trailing rate says 648 h

• ○ e3654f · Linux · x86-64 · no GPU · 32 GiB16 proved · 1624 queued · 14 it cannot reach

  away 2 h · last session ended without being asked to — power loss, a hard reset, or a crash

  expects 595 h for the rest (own samples) · trailing rate says 2599 h · blocked: fetch

• ○ afbb8a · macOS · ARM64 · Apple GPU · 16 GiB12 proved · 182 queued · 16 it cannot reach

  away 262 h · last session a supervisor bounce: a new session started while the last one was alive · this machine sleeps (46 min during its last session) — most likely asleep

  expects 23 h for the rest (own samples) · blocked: budget

• ○ 9e3c79 · Windows · x86-64 · Intel GPU · 16 GiB3 proved · 1337 queued · 22 it cannot reach

  away 18 min · last session a supervisor bounce: a new session started while the last one was alive

  expects 18 h for the rest (own samples) · blocked: budget

• ○ 0c5265 · macOS · ARM64 · Apple GPU · 8 GiB2 proved · 125 queued · 16 it cannot reach

  away 452 h · last session ended without being asked to — power loss, a hard reset, or a crash · this machine sleeps (37 min during its last session) and is on battery — most likely asleep

  expects 53 h for the rest (own samples) · trailing rate says 1500 h · blocked: fetch

• ○ d68672 · macOS · x86-64 · Apple GPU · 8 GiB1 proved · 999 queued · 37 it cannot reach

  away 32 min · last session a supervisor bounce: a new session started while the last one was alive

  expects 1076 h for the rest (own samples) · blocked: budget

• ○ e78fd2 · Linux · x86-64 · NVIDIA · 32 GiB1 proved · 139 queued · 55 it cannot reach

  away 188 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

• ○ 0548a8 · macOS · ARM64 · Apple GPU · 8 GiB0 proved · 144 queued · 55 it cannot reach

  away 133 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

• ○ 12954d · Windows · x86-64 · NVIDIA · 8 GiB0 proved · 1867 queued · 55 it cannot reach

  away 21 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

• ○ 2e5c02 · Windows · x86-64 · Intel GPU · 16 GiB0 proved · 127 queued · 55 it cannot reach

  away 88 h · last session a supervisor bounce: a new session started while the last one was alive · a laptop on battery — most likely asleep

  expects 7 h for the rest (own samples) · blocked: pace

• ○ 53b229 · Windows · x86-64 · NVIDIA · 16 GiB0 proved · 126 queued · 55 it cannot reach

  away 138 h · last session a supervisor bounce: a new session started while the last one was alive · this machine sleeps (10 min during its last session) and is on battery — most likely asleep

  expects 113 h for the rest (own samples) · blocked: pace

• ○ 84bf54 · Windows · x86-64 · NVIDIA · 16 GiB0 proved · 816 queued · 18 it cannot reach

  away 49 h · last session a supervisor bounce: a new session started while the last one was alive

  expects 42 h for the rest (own samples) · 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.