This is an experimental project built by a team of AI agents. The agents helped plan, build, review, and manage the project through GitHub, while I provided flexible instructions and overall direction.
Checked-in sources
OpenRouter coding/CLI-agent traffic
Bundled snapshot: data/openrouter-current.json
Bundled public snapshot
Source scope: Public OpenRouter apps; UTC window 2026-08-29 to 2026-09-27
Source date / last captured: 2026-09-28T17:22:56.744Z
Ranking metric: total tokens in the captured window; requests are shown as a separate count. These are not user counts.
Top 10 OpenRouter coding/CLI-agent apps by token volume
Rank
App
Tokens
Requests
1
Hermes Agent
48,113,350,542,381
494,002,814
2
Claude Code
22,673,023,764,970
234,043,808
3
Kilo Code
14,565,330,570,448
190,442,386
4
Cline
13,647,155,408,914
125,297,748
5
pi
9,011,655,797,722
88,490,185
6
omp
7,117,339,589,661
42,416,803
7
Codex
5,634,755,251,185
69,708,159
8
OpenClaw
5,192,857,041,197
74,752,543
9
OpenHands
1,931,541,157,440
20,224,117
10
Command Code
1,263,768,129,551
12,673,960
Six-month usage
Monthly public coding/CLI-agent top ten by tokens. An absent app has unknown counts, not zero; an unavailable snapshot is a separate error.
Hermes Agent: six completed UTC months
Month
Tokens
Requests
Top-list status
Captured as of
2026-03
341,465,597,590
6,385,676
returned rank 5
2026-09-28T20:19:21.577Z
2026-04
5,293,636,631,906
98,745,279
returned rank 3
2026-09-28T20:19:21.986Z
2026-05
11,855,069,268,842
199,365,362
returned rank 1
2026-09-28T20:19:22.987Z
2026-06
25,722,724,706,037
352,946,923
returned rank 1
2026-09-28T20:19:23.467Z
2026-07
29,688,895,154,398
363,476,983
returned rank 1
2026-09-28T20:19:23.901Z
2026-08
49,792,321,779,172
498,200,170
returned rank 1
2026-09-28T20:19:24.508Z
Source: OpenRouter App Rankings Data API; public aggregate data, CC BY 4.0. Captured `as_of` and window are shown above. Only returned top-ten apps are represented.
SWE-bench Verified results
Bundled snapshot: data/swe-bench-verified.json
Bundled public snapshot
Source scope: Submitted coding systems; five-record source sample, not a full leaderboard
Share of 1,340 professional respondents selecting one primary agent use case. Survey fielded 2025-11-18 to 2025-12-02. These survey percentages are not OpenRouter usage or app-specific measures.
Customer service
26.5%
Research & data analysis
24.4%
Internal workflow automation
18%
Text equivalent: published primary agent use-case shares
Metric: self-reported ever use of each named AI chatbot among U.S. adults. Pew surveyed U.S. adults Feb. 17–23, 2026.
These survey estimates are not active users, traffic, market share, OpenRouter usage, LangChain use cases, or SWE-bench performance. Brand questions used split forms and respondents could report more than one chatbot, so percentages must not be summed.
ChatGPT
44%
Gemini
24%
Copilot
17%
Meta AI
14%
Grok
8%
Claude
6%
Character.ai
3%
Text equivalent: self-reported ever use among U.S. adults
