Community Stats

What people run when every machine they own becomes one local AI endpoint. Live from the herd, last 30 days.

39
herds reporting this week
Herds per week 7 days to Sep 1: 4. 7 days to Sep 8: 6. 7 days to Sep 15: 5. 7 days to Sep 22: 9. 7 days to Sep 29: 10. 7 days to Oct 6: 14. 7 days to Oct 13: 20. 7 days to Oct 20: 19. 7 days to Oct 27: 24. 7 days to Nov 3: 22. 7 days to Nov 10: 31. 7 days to Nov 17: 39. 7 days to Sep 1: 4 herds7 days to Sep 8: 6 herds7 days to Sep 15: 5 herds7 days to Sep 22: 9 herds7 days to Sep 29: 10 herds7 days to Oct 6: 14 herds7 days to Oct 13: 20 herds7 days to Oct 20: 19 herds7 days to Oct 27: 24 herds7 days to Nov 3: 22 herds7 days to Nov 10: 31 herds7 days to Nov 17: 39 herds
Sep 1Nov 17
117
machines this week
Machines per week 7 days to Sep 1: 7. 7 days to Sep 8: 9. 7 days to Sep 15: 9. 7 days to Sep 22: 17. 7 days to Sep 29: 13. 7 days to Oct 6: 20. 7 days to Oct 13: 32. 7 days to Oct 20: 29. 7 days to Oct 27: 48. 7 days to Nov 3: 51. 7 days to Nov 10: 73. 7 days to Nov 17: 117. 7 days to Sep 1: 7 machines7 days to Sep 8: 9 machines7 days to Sep 15: 9 machines7 days to Sep 22: 17 machines7 days to Sep 29: 13 machines7 days to Oct 6: 20 machines7 days to Oct 13: 32 machines7 days to Oct 20: 29 machines7 days to Oct 27: 48 machines7 days to Nov 3: 51 machines7 days to Nov 10: 73 machines7 days to Nov 17: 117 machines
Sep 1Nov 17
72,581
requests routed this week
Requests routed per week 7 days to Sep 1: 52,980. 7 days to Sep 8: 61,179. 7 days to Sep 15: 60,463. 7 days to Sep 22: 66,376. 7 days to Sep 29: 52,256. 7 days to Oct 6: 94,437. 7 days to Oct 13: 76,027. 7 days to Oct 20: 62,200. 7 days to Oct 27: 71,281. 7 days to Nov 3: 71,098. 7 days to Nov 10: 80,416. 7 days to Nov 17: 72,581. 7 days to Sep 1: 52,980 requests7 days to Sep 8: 61,179 requests7 days to Sep 15: 60,463 requests7 days to Sep 22: 66,376 requests7 days to Sep 29: 52,256 requests7 days to Oct 6: 94,437 requests7 days to Oct 13: 76,027 requests7 days to Oct 20: 62,200 requests7 days to Oct 27: 71,281 requests7 days to Nov 3: 71,098 requests7 days to Nov 10: 80,416 requests7 days to Nov 17: 72,581 requests
Sep 1Nov 17
Weekly numbers
7 days toHerdsMachinesRequests routed
2026-11-173911772,581
2026-11-10317380,416
2026-11-03225171,098
2026-10-27244871,281
2026-10-20192962,200
2026-10-13203276,027
2026-10-06142094,437
2026-09-29101352,256
2026-09-2291766,376
2026-09-155960,463
2026-09-086961,179
2026-09-014752,980
  • 40 of 59 herds run on two or more machines.
  • The biggest herd pools 1,827 GB of memory across 13 machines.
  • The model the most herds run is qwen3-coder:30b, on 9 herds.
  • 534.8 million tokens in 30 days, on hardware people already own, with no API bill.

The totals

2026-10-19 to 2026-11-18.

Herds that routed requests54
Herds reporting in59
Machines in those herds158
Requests routed314,768
Prompt tokens463,396,349
Completion tokens71,381,591

How many machines are in a herd

How many machines are in a herd Number of herds by how many machines they reported. 1 machine: 19 herds. 2 machines: 24 herds. 3 machines: 4 herds. 4 machines: 4 herds. 5+ machines: 8 herds. 1 machine: 19 herds 1 machine 19 2 machines: 24 herds 2 machines 24 3 machines: 4 herds 3 machines 4 4 machines: 4 herds 4 machines 4 5+ machines: 8 herds 5+ machines 8
Herds by machine count, each machine counted once from its latest report in this window. Self-reported by herds with telemetry on; not verified.

Do bigger herds work harder?

Requests routed per machine in the last 30 days by the typical herd of each size.

Requests per machine by herd size Median requests routed per machine over 30 days, across the herds of each size. 1 machine · 19 herds: 24 requests. 2–3 machines · 28 herds: 44 requests. 4+ machines · 12 herds: 274 requests. 1 machine · 19 herds: 24 requests 1 machine · 19 herds 24 2–3 machines · 28 herds: 44 requests 2–3 machines · 28 herds 44 4+ machines · 12 herds: 274 requests 4+ machines · 12 herds 274
The median herd of each size: each herd's requests divided by its machines, idle herds included. A median, so one very busy herd can't move it. Self-reported; not verified.

Models the most herds run

Counted by herds, not requests. One busy herd can send most of the traffic in the request table below, so this is the better read on what is common.

Models by number of herds running them Number of herds that routed at least one request to each model. qwen3-coder:30b: 9 herds. nomic-embed-text: 7 herds. qwen3.5:9b: 7 herds. gemma4:e4b: 6 herds. gpt-oss:120b: 6 herds. gemma4:31b: 5 herds. llama3.2:3b: 5 herds. qwen2.5:7b: 5 herds. gemma3:27b: 4 herds. codestral:22b: 3 herds. qwen3-coder:30b: 9 herds qwen3-coder:30b 9 nomic-embed-text: 7 herds nomic-embed-text 7 qwen3.5:9b: 7 herds qwen3.5:9b 7 gemma4:e4b: 6 herds gemma4:e4b 6 gpt-oss:120b: 6 herds gpt-oss:120b 6 gemma4:31b: 5 herds gemma4:31b 5 llama3.2:3b: 5 herds llama3.2:3b 5 qwen2.5:7b: 5 herds qwen2.5:7b 5 gemma3:27b: 4 herds gemma3:27b 4 codestral:22b: 3 herds codestral:22b 3
Top 10 models by herds that routed at least one request to them, with "model:latest" counted as "model". Self-reported by herds with telemetry on; not verified.

Most-used models

#ModelRequestsShare
1gpt-oss:120b216,02269.77%
2dinov2-vit-s1437,06811.97%
3nomic-embed-text11,2023.62%
4qwen3.5:9b6,9122.23%
5llama3.2:1b6,1872.0%
6qwen3-coder:30b6,0471.95%
7gemma3:12b4,6691.51%
8gemma4:e4b4,6641.51%
9qwen3-embedding1,9810.64%
10qwen3:4b1,6560.53%
11qwen3.8:27b-mlx1,6100.52%
12qwen3.5-4b-sompoa-heresy-v2:q4_k_m1,5870.51%
13llama3.11,5230.49%
14qwen2.5-coder:14b1,4430.47%
15qwen3-embedding:0.6b1,4430.47%
16qwen3-coder1,2460.4%
17qwen3:32b9670.31%
18gemma4:12b7130.23%
19qwen2.5:7b5700.18%
20gemma4b:31-code-frame4190.14%
21qwen3.8:27b3520.11%
22gemma4:31b3450.11%
23qwen3.6-35b-a3b3400.11%
24gpt-oss:20b3230.1%
25gemma3:27b3130.1%

The herd leaderboard

Herds that report in, by name where the operator picked one and by a generated handle otherwise. Ranked by total memory, because that is more fun to argue about than request counts.

#HerdMachinesTotal memory
1 herd-0e7cf7 13 1827 GB
2 herd-44e41c 11 1577 GB
3 herd-9e82e7 4 896 GB
4 herd-9d2f64 7 686 GB
5 herd-dd0c06 6 666 GB

See the whole leaderboard →

Most common errors

#CategoryCountShare
1model_not_found1,51560.1%
2bad_request74629.59%
3context_too_long1566.19%
4other642.54%
5connection_error271.07%
6timeout70.28%
7server_error60.24%

Error categories are a fixed list. Raw error text never leaves anyone's machine.

About these numbers

Herds report a daily summary unless the operator turns it off with FLEET_NODE_TELEMETRY=false. Every field is listed on the telemetry page. The two herd counts differ on purpose: an idle herd reports its machines without routing any requests. Everything is self-reported, so read it as a strong signal, not a census.

Machine-readable

The same data is available as JSON at /api/v1/stats.