---
cursor:
  subagentId: "bc-63f48d87-11cd-551f-adb9-c617dedb2e99"
---

# Metrics sweep: agents layer (agents and tools; knowledge, search and security)

Access date for everything below: 24 Sep 2026. Full point lists are in `agents.json` (same folder); long daily series are subsampled in the tables here.

Subnets covered: Ridges 62, ORO 15, SOMA 114, Ditto 118, Gittensor 74, sundae_bar 121, Leadpoet 71, TrajectoryRL 11, RedTeam 61, Bitsec 60, ItsAI 32, Enigma 63, Trishool 23, Harnyx 67, Claims 111, Perturb 26, Aurelius 37, TAO Status 104. Desearch 22 and Apex 1 skipped (data worker).

Best finds: ORO has a fully public API (`api.oroagents.com/v1/public/...`) with a daily trajectories series since 26 Mar 2026; Ditto's landing-page counters can be read back through the Wayback Machine; Gittensor's `api.gittensor.io` exposes every PR and miner; SOMA's dashboard archives every competition round; sundae_bar files RNS announcements as a listed company.

## Series found

### oro-trajectories-cumulative

**ORO (SN15)** — Trajectories scored (distinct agent-task pairs), cumulative. Unit: trajectories; kind: cumulative; source type: chain; shape: **rising**; points: 183; first 2026-03-26 = 6; last 2026-09-24 = 4,176,132.

Public ORO API run by the team; daily since launch 26 Mar 2026; the same series is drawn on oroagents.com. Cumulative, so it rises by construction. Team posts quote a larger "judged trajectories" count (2,580,433 on 5 Jul 2026 vs 1.3M distinct here) because those count every validator run; this series counts distinct agent-and-task pairs. Emission-funded competition, no customers.

Subsampled (every 11 points; all 183 in JSON):

| date | value | verified | source |
|---|---|---|---|
| 2026-03-26 | 6 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-04-06 | 10,617 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-04-17 | 40,219 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-04-28 | 117,714 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-05-09 | 237,104 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-05-20 | 378,970 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-05-31 | 557,776 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-06-11 | 752,393 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-06-22 | 941,948 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-07-03 | 1,229,985 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-07-14 | 1,660,364 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-07-25 | 2,120,118 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-08-05 | 2,545,705 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-08-16 | 2,959,177 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-08-27 | 3,386,911 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-09-07 | 3,844,206 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-09-18 | 4,039,542 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-09-24 | 4,176,132 | yes | api.oroagents.com/v1/public/trajectories/history |

### oro-trajectories-per-week

**ORO (SN15)** — Trajectories scored per week (distinct agent-task pairs). Unit: trajectories/week; kind: rate; source type: chain; shape: **peaked**; points: 27; first 2026-03-23 = 1,222; last 2026-09-21 = 90,325.

Weekly sum of the daily API series; weeks start Monday. Peak 304,409 in week of 13 Jul 2026; fell to ~90K/week after the benchmark reset to ORO-Bench 3 (suites 140-144, Sep 2026). Last week (21 Sep) is partial (4 days).


| date | value | verified | source |
|---|---|---|---|
| 2026-03-23 | 1,222 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-03-30 | 7,389 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-04-06 | 10,577 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-04-13 | 30,989 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-04-20 | 49,929 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-04-27 | 59,849 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-05-04 | 89,328 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-05-11 | 94,964 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-05-18 | 103,316 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-05-25 | 110,213 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-06-01 | 120,968 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-06-08 | 120,909 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-06-15 | 124,826 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-06-22 | 157,567 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-06-29 | 210,219 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-07-06 | 279,548 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-07-13 | 304,409 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-07-20 | 282,693 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-07-27 | 275,648 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-08-03 | 262,949 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-08-10 | 261,665 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-08-17 | 270,491 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-08-24 | 271,279 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-08-31 | 299,302 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-09-07 | 172,961 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-09-14 | 112,597 | yes | api.oroagents.com/v1/public/trajectories/history |
| 2026-09-21 | 90,325 | yes | api.oroagents.com/v1/public/trajectories/history |

### oro-payouts-cumulative-usd

**ORO (SN15)** — Cumulative payouts to miners (USD, on-chain emissions). Unit: USD; kind: cumulative; source type: team; shape: **rising**; points: 4; first 2026-04-14 = 0; last 2026-08-04 = 1,000,000.

Team-stated in X posts; the team says the figure is Bittensor chain emissions valued in USD, not from the ORO API. Emission-funded, not customer revenue. Cumulative rises by construction. Launch 14 Apr 2026 (value 0).

| date | value | verified | source |
|---|---|---|---|
| 2026-04-14 | 0 | no | x.com/oroagents/status/2044208805118898360 |
| 2026-05-19 | 200,000 | no | x.com/oroagents/status/2057213959032524988 |
| 2026-07-05 | 738,000 | no | x.com/oroagents/status/2074627671607443891 |
| 2026-08-04 | 1,000,000 | no | x.com/oroagents/status/2084714798944752108 |

### oro-top-agent-daily-payout-usd

**ORO (SN15)** — Daily payout to the top agent. Unit: USD/day; kind: rate; source type: team; shape: **rising**; points: 4; first 2026-04-15 = 5,000; last 2026-09-24 = 21,310.

Prize per day at launch was $5,000 (team post 15 Apr 2026); "~$20K/day run-rate" (team image 5 Jul 2026); $21,310/day = 72.5 TAO at $293.77 from the public API top-miner-payout on 24 Sep 2026. Depends on TAO price and emission share; emission-funded.

| date | value | verified | source |
|---|---|---|---|
| 2026-04-15 | 5,000 | no | x.com/oroagents/status/2044208805118898360 |
| 2026-07-05 | 20,000 | no | x.com/oroagents/status/2074627671607443891 |
| 2026-07-07 | 20,000 | no | x.com/oroagents/status/2074627671607443891 |
| 2026-09-24 | 21,310 | yes | api.oroagents.com/v1/public/top-miner-payout |

### oro-qualifiers-per-race

**ORO (SN15)** — Agents qualifying per daily race. Unit: agents/race; kind: rate; source type: chain; shape: **peaked**; points: 158; first 2026-04-14 = 73; last 2026-09-23 = 231.

Public API races/history; one race a day since 14 Apr 2026 (161 races). Monthly means: Apr 144, May 273, Jun 368, Jul 391, Aug 354, Sep 282 (to 23 Sep). Peak 536 on 1 Jun 2026; dip after the benchmark reset in Sep 2026.

Subsampled (every 9 points; all 158 in JSON):

| date | value | verified | source |
|---|---|---|---|
| 2026-04-14 | 73 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-04-23 | 230 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-05-02 | 25 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-05-11 | 281 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-05-20 | 512 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-05-29 | 431 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-06-07 | 435 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-06-16 | 358 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-06-25 | 378 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-07-04 | 368 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-07-13 | 423 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-07-22 | 395 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-07-31 | 353 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-08-09 | 344 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-08-18 | 374 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-08-27 | 352 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-09-05 | 393 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-09-19 | 191 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |
| 2026-09-23 | 231 | yes | api.oroagents.com/v1/public/races/history?limit=50&offset=0 |

### oro-top-agent-score

**ORO (SN15)** — Score of the top agent when it took the lead. Unit: fraction of tasks solved; kind: rate; source type: chain; shape: **flat**; points: 134; first 2026-03-26 = 0.033; last 2026-09-24 = 0.7073.

Public API top/history. Problems change daily and suites reset (suite 1 Mar-Apr, suite 3 Apr-Sep, suites 140-144 Sep 2026), so scores are not comparable across resets; within suite 3 the peak was 0.878 on 21 Jul 2026 and later leaders scored 0.4-0.7. Not a clean growth series.

Subsampled (every 8 points; all 134 in JSON):

| date | value | verified | source |
|---|---|---|---|
| 2026-03-26 | 0.033 | yes | api.oroagents.com/v1/public/top/history |
| 2026-04-09 | 0.883 | yes | api.oroagents.com/v1/public/top/history |
| 2026-04-21 | 0.359 | yes | api.oroagents.com/v1/public/top/history |
| 2026-05-01 | 0.131 | yes | api.oroagents.com/v1/public/top/history |
| 2026-05-13 | 0.6 | yes | api.oroagents.com/v1/public/top/history |
| 2026-05-25 | 0.7227 | yes | api.oroagents.com/v1/public/top/history |
| 2026-06-03 | 0.4333 | yes | api.oroagents.com/v1/public/top/history |
| 2026-06-11 | 0.4 | yes | api.oroagents.com/v1/public/top/history |
| 2026-06-19 | 0.6167 | yes | api.oroagents.com/v1/public/top/history |
| 2026-06-28 | 0.5448 | yes | api.oroagents.com/v1/public/top/history |
| 2026-07-07 | 0.5056 | yes | api.oroagents.com/v1/public/top/history |
| 2026-07-17 | 0.7444 | yes | api.oroagents.com/v1/public/top/history |
| 2026-07-26 | 0.7667 | yes | api.oroagents.com/v1/public/top/history |
| 2026-08-08 | 0.5834 | yes | api.oroagents.com/v1/public/top/history |
| 2026-08-22 | 0.5333 | yes | api.oroagents.com/v1/public/top/history |
| 2026-09-04 | 0.5667 | yes | api.oroagents.com/v1/public/top/history |
| 2026-09-18 | 0.6282 | yes | api.oroagents.com/v1/public/top/history |
| 2026-09-24 | 0.7073 | yes | api.oroagents.com/v1/public/top/history |

### oro-top50-agent-score-weekly

**ORO (SN15)** — Mean score of the top 50% of agents, by week (first 5 weeks). Unit: percent; kind: rate; source type: team; shape: **U**; points: 5; first 2026-04-20 = 49.1; last 2026-05-18 = 58.9.

From the team image "36 days of ORO races" (weeks 1-5 from 14 Apr 2026); team says +11.5% week over week after week 2. Dates are week ends. Only 5 weeks; not continued publicly.

| date | value | verified | source |
|---|---|---|---|
| 2026-04-20 | 49.1 | no | x.com/oroagents/status/2057213959032524988 |
| 2026-04-27 | 42.7 | no | x.com/oroagents/status/2057213959032524988 |
| 2026-05-04 | 43.9 | no | x.com/oroagents/status/2057213959032524988 |
| 2026-05-11 | 51.2 | no | x.com/oroagents/status/2057213959032524988 |
| 2026-05-18 | 58.9 | no | x.com/oroagents/status/2057213959032524988 |

### oro-agent-versions-submitted-cumulative

**ORO (SN15)** — Agent versions submitted, cumulative. Unit: agents; kind: cumulative; source type: team; shape: **rising**; points: 3; first 2026-05-14 = 4,000; last 2026-07-05 = 16,786.

Team posts and images: "nearing 4000" (14 May), 5,754 agents (14 Apr-19 May image), 16,786 distinct agent versions evaluated (5 Jul image). Current suite alone shows 516 competing agents and 240 submitted in 24h (API, 24 Sep 2026). Cumulative.

| date | value | verified | source |
|---|---|---|---|
| 2026-05-14 | 4,000 | no | x.com/oroagents/status/2054985113612722559 |
| 2026-05-19 | 5,754 | no | x.com/oroagents/status/2057213959032524988 |
| 2026-07-05 | 16,786 | no | x.com/oroagents/status/2074627671607443891 |

### ditto-users

**Ditto (SN118)** — Users (landing-page counter). Unit: users; kind: cumulative; source type: team; shape: **rising**; points: 4; first 2026-05-08 = 790; last 2026-09-24 = 2,500.

Counter on heyditto.ai ("Users 2,500+"); earlier values read from Wayback Machine snapshots of the same page. Values are rounded "N+" figures. Team-run page, self-reported.

| date | value | verified | source |
|---|---|---|---|
| 2026-05-08 | 790 | yes | web.archive.org/web/20260508125847/heyditto.ai/ |
| 2026-06-11 | 1,100 | yes | web.archive.org/web/20260611010944/heyditto.ai/?ref=tao.media |
| 2026-09-15 | 2,500 | yes | web.archive.org/web/20260915150117/heyditto.ai/ |
| 2026-09-24 | 2,500 | yes | heyditto.ai/ |

### ditto-prompts

**Ditto (SN118)** — Prompts (landing-page counter). Unit: prompts; kind: cumulative; source type: team; shape: **rising**; points: 4; first 2026-05-08 = 38,400; last 2026-09-24 = 58,000.

Same page and snapshots as ditto-users; labelled "conversations remembered" in May 2026 and "Prompts" from June. Cumulative.

| date | value | verified | source |
|---|---|---|---|
| 2026-05-08 | 38,400 | yes | web.archive.org/web/20260508125847/heyditto.ai/ |
| 2026-06-11 | 46,300 | yes | web.archive.org/web/20260611010944/heyditto.ai/?ref=tao.media |
| 2026-09-15 | 58,000 | yes | web.archive.org/web/20260915150117/heyditto.ai/ |
| 2026-09-24 | 58,000 | yes | heyditto.ai/ |

### gittensor-merged-prs-per-week-sparkinfer

**Gittensor (SN74)** — Pull requests merged per week into the whitelisted repos (sparkinfer + spark-hermes). Unit: PRs/week; kind: rate; source type: chain; shape: **flat**; points: 14; first 2026-06-22 = 23; last 2026-09-21 = 9.

From api.gittensor.io/prs (779 PRs since 23 Jun 2026; 330 merged). Only the two currently whitelisted repos are in this API; earlier repos (Phase, vanguarstew, gittensor-ui) are gone from it, so this understates the network. Weeks start Monday; last week partial.

| date | value | verified | source |
|---|---|---|---|
| 2026-06-22 | 23 | yes | api.gittensor.io/prs |
| 2026-06-29 | 15 | yes | api.gittensor.io/prs |
| 2026-07-06 | 16 | yes | api.gittensor.io/prs |
| 2026-07-13 | 17 | yes | api.gittensor.io/prs |
| 2026-07-20 | 19 | yes | api.gittensor.io/prs |
| 2026-07-27 | 8 | yes | api.gittensor.io/prs |
| 2026-08-03 | 23 | yes | api.gittensor.io/prs |
| 2026-08-10 | 62 | yes | api.gittensor.io/prs |
| 2026-08-17 | 21 | yes | api.gittensor.io/prs |
| 2026-08-24 | 10 | yes | api.gittensor.io/prs |
| 2026-08-31 | 29 | yes | api.gittensor.io/prs |
| 2026-09-07 | 50 | yes | api.gittensor.io/prs |
| 2026-09-14 | 28 | yes | api.gittensor.io/prs |
| 2026-09-21 | 9 | yes | api.gittensor.io/prs |

### gittensor-merged-prs-cumulative-sparkinfer

**Gittensor (SN74)** — Pull requests merged into sparkinfer/spark-hermes, cumulative. Unit: PRs; kind: cumulative; source type: chain; shape: **rising**; points: 14; first 2026-06-22 = 23; last 2026-09-21 = 330.

Cumulative form of the weekly series; rises by construction. 56,888 lines added and 7,228 deleted across the 330 merged PRs.

| date | value | verified | source |
|---|---|---|---|
| 2026-06-22 | 23 | yes | api.gittensor.io/prs |
| 2026-06-29 | 38 | yes | api.gittensor.io/prs |
| 2026-07-06 | 54 | yes | api.gittensor.io/prs |
| 2026-07-13 | 71 | yes | api.gittensor.io/prs |
| 2026-07-20 | 90 | yes | api.gittensor.io/prs |
| 2026-07-27 | 98 | yes | api.gittensor.io/prs |
| 2026-08-03 | 121 | yes | api.gittensor.io/prs |
| 2026-08-10 | 183 | yes | api.gittensor.io/prs |
| 2026-08-17 | 204 | yes | api.gittensor.io/prs |
| 2026-08-24 | 214 | yes | api.gittensor.io/prs |
| 2026-08-31 | 243 | yes | api.gittensor.io/prs |
| 2026-09-07 | 293 | yes | api.gittensor.io/prs |
| 2026-09-14 | 321 | yes | api.gittensor.io/prs |
| 2026-09-21 | 330 | yes | api.gittensor.io/prs |

### gittensor-record-prs-merged-in-a-day

**Gittensor (SN74)** — Record for most PRs merged in a single day (across all whitelisted repos). Unit: PRs/day; kind: rate; source type: team; shape: **rising**; points: 7; first 2026-05-29 = 78; last 2026-06-23 = 1,130.

Team post 24 Jun 2026 listing record days; the 78 held from 1 Jan to 29 May 2026 (dated here at 29 May). Records only increase by construction; the jump came with the first "agentic maintained repo" (Phase). The attached chart shows 8.1k merged PRs all-time, 2.7k issues resolved, 12k PRs opened.

| date | value | verified | source |
|---|---|---|---|
| 2026-05-29 | 78 | no | x.com/gittensor_io/status/2069816961354109101 |
| 2026-05-31 | 242 | no | x.com/gittensor_io/status/2069816961354109101 |
| 2026-06-03 | 242 | no | x.com/gittensor_io/status/2069816961354109101 |
| 2026-06-05 | 271 | no | x.com/gittensor_io/status/2069816961354109101 |
| 2026-06-18 | 636 | no | x.com/gittensor_io/status/2069816961354109101 |
| 2026-06-20 | 1,126 | no | x.com/gittensor_io/status/2069816961354109101 |
| 2026-06-23 | 1,130 | no | x.com/gittensor_io/status/2069816961354109101 |

### soma-uploads-per-competition

**SOMA (SN114)** — Model uploads per competition round (CoT-Compression 2-10). Unit: uploads/round; kind: rate; source type: chain; shape: **peaked**; points: 9; first 2026-06-03 = 195; last 2026-09-29 = 20.

Team dashboard thesoma.ai/dashboard?comp=<id>; dated at upload close. Rounds 1-6 (Context-Compression) and CoT-1 show 0 uploads in the archive. Peak 555 in CoT-6 (28 Jul 2026); CoT-10 (20) is still open to 29 Sep. Prize pool fixed at 715.5 alpha per round.

| date | value | verified | source |
|---|---|---|---|
| 2026-06-03 | 195 | yes | thesoma.ai/dashboard?comp=75 |
| 2026-06-17 | 227 | yes | thesoma.ai/dashboard?comp=107 |
| 2026-07-01 | 324 | yes | thesoma.ai/dashboard?comp=108 |
| 2026-07-14 | 492 | yes | thesoma.ai/dashboard?comp=110 |
| 2026-07-28 | 555 | yes | thesoma.ai/dashboard?comp=111 |
| 2026-08-11 | 458 | yes | thesoma.ai/dashboard?comp=112 |
| 2026-08-27 | 236 | yes | thesoma.ai/dashboard?comp=145 |
| 2026-09-13 | 120 | yes | thesoma.ai/dashboard?comp=146 |
| 2026-09-29 | 20 | yes | thesoma.ai/dashboard?comp=147 |

### sundaebar-marketplace-listings

**sundae_bar (SN121)** — AI agents / skills listed on the sundaebar.ai marketplace. Unit: listings; kind: cumulative; source type: team; shape: **rising**; points: 3; first 2025-10-15 = 174; last 2026-09-24 = 700.

From RNS regulatory announcements: 174 AI agents (payments-live RNS, 15 Oct 2025); "more than 300 AI agents and workflows" (interim results, 29 May 2026); "more than 700 Skills, Agents and other capabilities" (subscription launch RNS, 24 Sep 2026). The unit widened from agents to skills+agents over time.

| date | value | verified | source |
|---|---|---|---|
| 2025-10-15 | 174 | no | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/payments-live-on-sundae_bar-platform/9171361 |
| 2026-05-29 | 300 | no | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/interim-results-for-six-months-ended-31-march-2026/9586461 |
| 2026-09-24 | 700 | no | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/launch-of-ai-agent-subscription-on-sundae_bar/9788010 |

### sundaebar-revenue-half-year

**sundae_bar (SN121)** — Revenue per half-year (GBP, listed company accounts). Unit: GBP; kind: rate; source type: team; shape: **rising**; points: 3; first 2025-03-31 = 0; last 2026-03-31 = 7,944.

AIM-listed Sundae Bar AI PLC. H1 FY25 (to 31 Mar 2025): "generated no revenue". H2 FY25 (to 30 Sep 2025): nil, first revenue was recognised Mar 2026. H1 FY26 (to 31 Mar 2026): revenue £7,944 plus "other operating income" £711,948, which is mostly SN121 alpha-token emissions, not customer revenue. Rising from zero but tiny; the £719,892 headline mixes both.

| date | value | verified | source |
|---|---|---|---|
| 2025-03-31 | 0 | no | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/interim-results/8952992 |
| 2025-09-30 | 0 | no | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/final-results-for-the-year-ended-30-september-2025/9492551 |
| 2026-03-31 | 7,944 | no | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/interim-results-for-six-months-ended-31-march-2026/9586461 |

### ridges-polyglot-python-score

**Ridges (SN62)** — Top agent score on Polyglot Python (with test patch). Unit: percent; kind: rate; source type: team; shape: **flat**; points: 3; first 2025-12-16 = 96; last 2026-03-17 = 96.3.

Same benchmark quoted three times: ~96% on 135 problems (TAO Daily, 16 Dec 2025), "over 96%" (team post 21 Dec 2025), 96.3% (team post 17 Mar 2026). Saturated; later posts use harder hidden sets (76.67% comp 22) that are not comparable.

| date | value | verified | source |
|---|---|---|---|
| 2025-12-16 | 96 | no | taodaily.io/ridges-ai-sets-a-new-standard-for-open-source-coding-agents/ |
| 2025-12-21 | 96 | no | x.com/ridges_ai/status/2002676651524829211 |
| 2026-03-17 | 96.3 | no | x.com/ridges_ai/status/2033921482217427382 |

## Single numbers

| subnet | netuid | metric | value | unit | date | source | type | note |
|---|---|---|---|---|---|---|---|---|
| ORO | 15 | Unique miners that submitted (first 36 days) | 1,206 | miners | 2026-05-19 | x.com/oroagents/status/2057213959032524988 | team | 5,754 agents from 1,206 miners, 14 Apr-19 May 2026; 29 unique race winners |
| ORO | 15 | Registered miners (builders), cumulative | 2,577 | miners | 2026-07-05 | x.com/oroagents/status/2074627671607443891 | team | Image "traction at a glance", as-of 5 Jul 2026 |
| ORO | 15 | Unique miners in current suite (ORO-Bench 3) | 310 | miners | 2026-09-24 | api.oroagents.com/v1/public/leaderboard?limit=100&offset=0&score_type=race | chain | 516 competing agents, 240 submitted in last 24h |
| ORO | 15 | Agents submitted in last 24h | 240 | agents/day | 2026-09-24 | api.oroagents.com/v1/public/leaderboard?limit=100&offset=0&score_type=race | chain | rate; current suite |
| ORO | 15 | Races run (one per day) | 161 | races | 2026-09-24 | api.oroagents.com/v1/public/races/history?limit=50&offset=0 | chain | 14 Apr-23 Sep 2026; team said 140 competitions on 4 Aug |
| ORO | 15 | Builders rewarded on-chain / top-agent awards | 83 | builders | 2026-07-05 | x.com/oroagents/status/2074627671607443891 | team | 97 top-agent awards to date |
| ORO | 15 | Judged trajectories, all-time (team count incl. repeat runs) | 2,580,433 | trajectories | 2026-07-05 | x.com/oroagents/status/2074627671607443891 | team | ~59K/day in the prior 7 days; 51% strict pass rate (1.32M of 2.58M); weekly chart 17K to 424K/week Mar 30-Jun 29 |
| ORO | 15 | Trajectories per day (team-stated) | 20,000 | trajectories/day | 2026-06-01 | x.com/oroagents/status/2061441585535918535 | team | "20k high-quality trajectories per day" after 50 days; API distinct-pairs rate that week ~17K/day |
| ORO | 15 | Post-training lift on Qwen3-4B (shopping tasks) | 42.7 | percent | 2026-07-07 | x.com/oroagents/status/2074627671607443891 | team | 18% base to 42.7% (53.3% pass@8); Sonnet 4.6 at 64%; arena top agent 77.3% on the 75-problem held-out set |
| ORO | 15 | Products in benchmark catalogue | 2,500,000 | products | 2026-04-20 | docs.oroagents.com | team | "2.5M real products" per team; current suite catalog_epoch lazada-schema-b-full-2026-08-31, 930 tasks (API suites/current) |
| ORO | 15 | Peak request load absorbed | 500,000,000 | requests/day | 2026-06-15 | x.com/shardiban/status/2066654068039111153 | team | "DDoSed with 500 million requests in a day", platform stayed up; 0 days downtime claimed through 4 Aug |
| ORO | 15 | Top-miner payout run-rate | 72.54 | TAO/day | 2026-09-24 | api.oroagents.com/v1/public/top-miner-payout | chain | = $21,310/day at $293.77 TAO |
| ORO | 15 | Y Combinator admission | 1 | event | 2026-09-08 | x.com/oroagents/status/2097396633764032707 | team | YC Fall 2026 batch |
| Ridges | 62 | Competition 22 agents created | 419 | agents | 2026-05-22 | x.com/ridges_ai/status/2057823982578225231 | team | 14-19 May 2026: 405 passed pre-screening, 155 beyond screener 1, 112 completed validators, 9 approved; 236 unique miners; best 76.67%, avg 57.56% |
| Ridges | 62 | Average cost per problem solved | 0.1 | USD/problem | 2026-05-22 | x.com/ridges_ai/status/2057823982578225231 | team | cost cap $0.29; 65% margin claimed |
| Ridges | 62 | SWE-Bench Verified score range | 87.8 | percent | 2026-03-17 | x.com/ridges_ai/status/2033921482217427382 | team | "73.6-87.8% on SWE-Bench Verified", $29/month for ~100 PRs claimed at the time |
| Ridges | 62 | Linting niche: hidden-test score after one week | 76 | percent | 2026-08-26 | x.com/ridges_ai/status/2092684102595961119 | team | cost $0.02 per task |
| Ridges | 62 | Evaluations per day (capacity) | 1,000 | evals/day | 2026-01-12 | x.com/ridges_ai/status/2010846945217003730 | team | "scaling evals beyond 1k per day" |
| Ridges | 62 | Ridgeline pricing | 9.99 | USD/month | 2026-07-21 | x.com/ridges_ai/status/2079567082874871969 | team | 10 credits (one issue each); free for stakers; x402 pay-per-credit $1.09 (app.ridges.ai, 24 Sep 2026). No user or issue counts published; ridges.ai and platform-v2.ridges.ai are behind a Vercel challenge. |
| Ridges | 62 | X followers | 8,119 | followers | 2026-09-24 | api.fxtwitter.com/ridges_ai | third-party | 735 posts |
| Gittensor | 74 | Total lines changed (all-time, network) | 394,279 | lines | 2026-09-24 | api.gittensor.io/dash/stats | chain | also 48 unique repositories, 15,941 issues, 1,326 commits; TAO Daily quoted "over 200,000 lines of production code" on 4 Feb 2026 (2 points only) |
| Gittensor | 74 | Lines of production code (team via TAO Daily) | 200,000 | lines | 2026-02-04 | taodaily.io/gittensor-pays-developers-for-contributing-to-open-source-projects-on-bittensor/ | third-party | "over 200,000" |
| Gittensor | 74 | Merged PRs all-time (dashboard chart) | 8,100 | PRs | 2026-06-24 | x.com/gittensor_io/status/2069816961354109101 | team | chart legend: Merged PRs 8.1k, Issues Resolved 2.7k, PRs Opened 12k, Issues Opened 5.3k |
| Gittensor | 74 | Merged PRs summed over currently registered miners | 5,789 | PRs | 2026-09-24 | api.gittensor.io/miners | chain | 180 registered miners, 108 with a merged PR, 9,617 PRs total; deregistered miners drop out so this is below the 8.1k all-time |
| Gittensor | 74 | Monthly reward pool (estimate) | 125,258 | USD/month | 2026-02-04 | taodaily.io/gittensor-pays-developers-for-contributing-to-open-source-projects-on-bittensor/ | third-party | Now ~$307/day (sparkinfer) + ~$102/day (spark-hermes) ≈ $12K/month on gittensor.io, 24 Sep 2026: falling, emission share ~0% under the price-based model |
| Gittensor | 74 | Reward rate shown on site | 409 | USD/day | 2026-09-24 | gittensor.io/ | chain | sparkinfer ~$307/day, 17 active miners, 24 PRs merged this week; spark-hermes ~$102/day, 7 miners, 5 PRs |
| Gittensor | 74 | PRs merged, early 2026 rate | 131 | PRs | 2026-02-04 | taodaily.io/gittensor-pays-developers-for-contributing-to-open-source-projects-on-bittensor/ | third-party | "from 60 pull requests in 10 days (12 Jan) to 131 in 13 days" |
| Gittensor | 74 | Phase repo, June 2026: merged PRs in the month | 1,783 | PRs/month | 2026-07-01 | x.com/gittensor_io/status/2072436835448770603 | team | 67 authors, 1,827 commits to main; one repo |
| Gittensor | 74 | Qwen3.8 decode speedup from one merged PR | 106.4 | tok/s | 2026-08-21 | x.com/gittensor_io/status/2090620190102782176 | team | 90.04 to 106.40 tok/s on RTX 5090, +18.2% |
| SOMA | 114 | Prize pool per competition round | 715.5 | alpha | 2026-09-24 | thesoma.ai/dashboard | chain | 41,328 in alpha shown; 3 validators (Rizzo, SOMA, Yuma); registration 0.25 alpha; no token-routing or savings counter exists on the site |
| SOMA | 114 | Compression price discount on gateway | 15 | percent | 2026-09-24 | thesoma.ai/ | team | DeepSeek V4 Pro $0.92 vs $0.78 per 1M input tokens (was ~10% on 17 Sep) |
| sundae_bar | 121 | Revenue and other operating income, H1 FY26 | 719,892 | GBP | 2026-05-29 | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/interim-results-for-six-months-ended-31-march-2026/9586461 | team | revenue £7,944 + other operating income £711,948 (largely SN121 alpha); operating loss £575,248 |
| sundae_bar | 121 | First enterprise contract value | 31,250 | USD | 2026-03-17 | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/commercial-contract-for-ai-agent-deployment/9477028 | team | AI agents for operational workflows |
| sundae_bar | 121 | SN121 developer submissions | 1,600 | submissions | 2026-06-01 | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/sundae_bar-ai-agent-platform-update-name-change/9707349 | team | "1,600+ SN121 submissions" by June 2026 (month only) |
| sundae_bar | 121 | Skills directory size at launch | 120 | skills | 2026-04-01 | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/sundae_bar-ai-agent-platform-update-name-change/9707349 | team | April 2026 (month only); 700+ skills/agents by 24 Sep 2026 |
| sundae_bar | 121 | Legacy user base (Ora + Kondor) | 60,000 | users | 2025-08-20 | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/launch-of-sundae_bar-platform-version-1-0/9064381 | team | "community of over 60,000 users" from the merged companies, not marketplace users |
| sundae_bar | 121 | Subscription price | 25 | GBP/month | 2026-09-24 | www.investegate.co.uk/announcement/rns/sundae-bar-plc--sbar/launch-of-ai-agent-subscription-on-sundae_bar/9788010 | team | 9 pre-built agents + Scout with 700+ skills |
| Bitsec | 60 | Agents submitted, cumulative (max agent id) | 5,178 | agents | 2026-09-24 | bitsec.ai/api/jobs/leaderboard | chain | approximate: highest agent_id on the leaderboard; 120 agents in round v3.1.11 (closed 21 Sep 2026), 119 submitted on 21 Sep |
| Bitsec | 60 | Agents created last week | 120 | agents/week | 2026-09-24 | bitsec.ai/api/jobs/stats | chain | real_vulns_found 0 for the current project set; daily prize pool $1,000; 5 validators; 4 public + 3 hidden projects. No scan or subscriber counters. |
| RedTeam | 61 | Annualized revenue (Innerworks) | 1,000,000 | USD | 2026-08-11 | taodaily.io/how-innerworks-built-redteam-sn61-into-a-seven-figure-business/ | third-party | "seven-figure annualized revenue" in six months, 90%+ gross margins, 50/50 fiat vs crypto, ~20 staff; CEO statement on Revenue Search podcast; exact figure not given; no attack/detection counts published |
| Claims | 111 | Raised at Proof of Pitch (Louvre, Paris) | 525 | TAO | 2026-06-01 | claims111.ai/ | team | 184 TAO pledged live on stage; June 2026 (month only); no claims-extracted counter on site |
| Aurelius | 37 | Public dataset on Hugging Face | 1 | datasets | 2026-06-11 | huggingface.co/api/datasets/AureliusAligned/gemma3-12b-baseline-pool | third-party | gemma3-12b-baseline-pool, 10K-100K rows, 3 downloads, 2 commits both 11 Jun 2026; no dataset-size series |
| TAO Status | 104 | Benchmark runs recorded (gpt-4o-mini) | 10 | runs | 2026-09-24 | taostatus.com/api/galileo/dashboard?window=30&include_scheduled=true | chain | all-time avg 18.8; gemini-3.6-flash 1 run (90), gpt-4o 1 run (52). Site is a "Bittensor Arena" beta; no subnet catalogue counter. |
| Ditto | 118 | X followers | 1,644 | followers | 2026-09-24 | api.fxtwitter.com/heydittoai | third-party | 951 posts |

## Not rising

- `oro-trajectories-per-week` — peaked (304K in week of 13 Jul 2026, ~90K/week now after the ORO-Bench 3 reset; last week partial).
- `oro-qualifiers-per-race` — peaked (monthly mean 391 in Jul, 282 in Sep after the suite reset).
- `oro-top-agent-score` — flat/noisy; scores are not comparable across suite resets.
- `oro-top50-agent-score-weekly` — U (49.1 → 42.7 → 58.9 over five weeks; series stops 19 May 2026).
- `gittensor-merged-prs-per-week-sparkinfer` — flat/noisy (8 to 62 PRs a week); only the current two repos.
- `soma-uploads-per-competition` — peaked (555 in CoT-6, 120 in CoT-9).
- `ridges-polyglot-python-score` — flat at ~96% (saturated benchmark).
- `sundaebar-revenue-half-year` — rising only from zero to £7,944; the £719,892 headline is mostly alpha-token income.
- Gittensor monthly reward pool (two points, not a series): $125,258/month (Feb 2026) → ~$12K/month (Sep 2026), falling.

## Could not find

- Ridges SN62: top-agent score over time on SWE-bench/Polyglot as a dated series (only 3 saturated Polyglot points; ridges.ai leaderboard and platform-v2.ridges.ai return a Vercel security challenge, no Wayback captures; competition-results posts are not regular: only comp 22 found in X search; app.ridges.ai has no counters); Ridgeline issues solved; users; Harbor evaluation counts.
- ORO SN15: dated cumulative payout points beyond $200K (19 May), ~$738K (5 Jul), $1M (4 Aug); a unique-miners series (only 1,206 and 2,577); products-in-catalogue over time (only the 2.5M claim).
- SOMA SN114: tokens routed or dollars saved (no counter on thesoma.ai or the dashboard; app.thesoma.ai needs login).
- Gittensor SN74: monthly reward pool over time (two points only; taodaily Feb 2026 and gittensor.io today); all-time merged-PR series before 23 Jun 2026 (api.gittensor.io/prs only holds the current repos; no Wayback captures of gittensor.io).
- sundae_bar SN121: revenue table for FY25 (investegate shows the summary only; revenue was nil per the RNS text); 500+ listings claim (Aug 2026) not in the Aug RNS text.
- Leadpoet SN71: no counters on leadpoet.com (site timed out under Playwright; static text only).
- TrajectoryRL SN11: no website counters; X account 114 followers, 19 posts.
- RedTeam SN61 / Innerworks: customers, attacks or detections counts (innerworks.me and theredteam.io have no counters; no official X account).
- Bitsec SN60: scans, vulnerabilities found, subscribers (API shows real_vulns_found 0 for the current set; no history endpoint; project_set_id filter ignored).
- ItsAI SN32: detections, API calls, monthly visits (its-ai.org shows pricing and accuracy claims only; api.its-ai.org has no public stats route; the ~22k monthly visits figure could not be sourced).
- Enigma SN63: bounties (no website found; EnigmaSN63 has no site link; qBitTensor GitHub org not found).
- Trishool SN23, Harnyx SN67, Perturb SN26: no usage counters on their sites (Harnyx shows comparator pricing; Perturb shows demo robustness percentages).
- Claims SN111: claims-extracted count (claims111.ai has no counter; GitHub DeSciClaims/Claims 5 stars).
- Aurelius SN37: dataset size over time (one HF dataset with 2 commits on the same day; no other org datasets).
- TAO Status SN104: a subnet catalogue count (taostatus.com is an LLM-eval arena beta; API only exposes model scores).
- Agent-marketplace counters beyond sundae_bar (none found).

## X reads used

28 of 30 (4 calls, logged in `/tmp/sweep/xreads-agents.log`): `from:ridges_ai Competition Results` (2), `from:ridges_ai (record OR "top agent" OR ...)` (9), `(from:oroagents OR from:shardiban) (paid OR payouts ...)` (7), `from:gittensor_io ("lines of code" OR merged OR ...)` (10).

## Charts saved

All in `media/charts-gallery/` in the Project store.

| file | source URL | what it shows |
|---|---|---|
| oro-trajectories-cumulative.png | https://oroagents.com/ (figure "Trajectories scored", screenshot 24 Sep 2026) | cumulative distinct trajectories 26 Mar–24 Sep 2026, 4,176,066 |
| oro-traction-dashboard-jul-2026.png | https://x.com/oroagents/status/2074627671607443891 (as-of 5 Jul 2026) | 2,580,433 judged trajectories, ~$738K paid, 16,786 agent versions, 2,577 miners, weekly trajectories 17K→424K |
| oro-36-days-submissions.jpg | https://x.com/oroagents/status/2057213959032524988 | daily submissions 14 Apr–19 May 2026, 5,754 agents, 1,206 miners, $200K+ payout, top-50% score by week |
| oro-top-agent-score-per-race.png | https://oroagents.com/leaderboard ("Challenge History", screenshot 24 Sep 2026) | top-agent score per race Apr–Sep 2026 |
| gittensor-merged-prs-active-network.png | https://x.com/gittensor_io/status/2069816961354109101 | merged PRs per day Nov 2025–Jun 2026 (8.1k all-time) |
| ridges-competition-22-funnel.jpg | https://x.com/ridges_ai/status/2057823982578225231 | competition 22 funnel: 419 agents → 9 approved, 76.67% best |