Benchmark: search events on Eventim with an AI agent
On Eventim, Wield's recipe passed 9 of 9 event searches with a median of 286 ms and no model cost per run. A browser agent running Claude Sonnet 5.5 passed 0 of 9 runs, with a median of 121 seconds and $0.06 per run. The benchmark counts all 9 agent runs as blocked by the site. The agent reported the site unavailable, and its log shows no HTTP error status.
Eventim Search concerts and events with dates, venues and ticket availability
Read-only, no loginRuns in your browser only
Parameters (5)
| Metric | Wield recipe | Browser Use + Claude Sonnet 5.5 |
|---|---|---|
| Success | 100% | 0% |
| Median time | 286 ms | 121 s |
| Cost per run | $0 | $0.065 |
9 runs each, 7 November 2026. How we measure
The task
Both sides ran the same three searches.
| Query | Search text |
|---|---|
| q1 | Die Ärzte |
| q2 | Helene Fischer |
| q3 | Disney on Ice |
Pass check: at least three events, each with a name and a date.
Results (7 Nov 2026)
| Wield recipe | Browser agent | |
|---|---|---|
| Runs passing the check | 9 of 9 | 0 of 9 |
| Median time per run | 286 ms | 120.9 s |
| Model cost per run | $0 | $0.06 |
| Requests per run (median) | 1 | 14 (0 to the site) |
| Blocked by the site | 0 runs | 9 of 9 runs |
The browser agent took a median of 6 steps, 14,196 input tokens and 1,018 output tokens per run. Its log records no HTTP 401, 403 or 429 response from the site, so the block does not show as a status code.
Wield's raw logs also hold one entry that stopped with origin_blocked. The browser had not granted access to https://www.eventim.de. The table uses the summary counts, which show 9 of 9 passing. Wield's fastest run took 169 ms, and its slowest took 120,270 ms, so the median does not show every run.
Why the results differ
Eventim's public event API returns event groups, and the recipe reads it from a www.eventim.de tab without sending credentials. The API carries no prices, so the recipe returns event names and dates, not ticket prices.
The browser agent loads the Eventim site in a fresh headless browser. The agent reported the site as unavailable in its runs. Wield's recipe does not load the site in that way. It makes one request from a tab in your own browser.
Caveats
- Small and preliminary. The benchmark was still running on 7 Nov 2026. Pacing at about 30 site requests per hour keeps counts low on some sides.
- A block, not a model result. The 0 of 9 comes from the site's behaviour toward the agent's browser. It says nothing about the agent's search skill on other sites.
- Cost is model cost only. Wield's $0 covers model cost. No model runs in the loop when a recipe runs.
Terms of Eventim
Eventim's terms prohibit automated access to its systems. The recipe runs one search at a time that you start, from your own browser, with one request. Buying tickets stays on eventim.de. Wield's recipe does not automate queueing or purchases. Check the site's terms yourself. Wield's recipes note known restrictions.
FAQ
Does Eventim block browser agents?
In these runs, yes. The benchmark counts 9 of 9 browser agent runs as blocked by the site. The agent reported the site unavailable. Its log shows no 401, 403 or 429 status, so the log does not show the cause. Results may differ on other days, and bot protection can change at any time.
How fast is the Wield recipe for search events on Eventim?
The median run took 286 ms over 9 runs. The fastest took 169 ms, and the slowest took 120,270 ms. All 9 runs passed the check in the summary.
Can I rerun this benchmark?
Yes. The method is on the benchmark page, and the query table above is the full input. The raw per-run logs keep destinations, timing, tokens and checks. They do not keep page contents.