B0 Orchestration floor
The URL goes straight to the output node. No fetch, no conversion, no model.
On the other pages this cell is B0 URL → output, nothing runs. Its results across models.
What each of the ten benchmark cells does, drawn with one shared vocabulary, and how the cells relate: which pairs differ by exactly one decision.
Every cell runs the same prompt on the same pages. What differs is the architecture around the model: what it is given (a URL, the HTML, or Markdown), whether it runs once or in a loop, which tools it holds, and who owns them. The labels used everywhere on this site read input the model sees → who does the work → tools.
The floor and the reference. Free, deterministic, and the yardstick for time and structure.
A fixed graph with a single model call in it. The graph decides the order of work; the model does one job.
The model decides what runs and when. What moves between cells is the input it is given, the tools it holds, and who owns them.
Each pair differs by one decision, so the gap between the two rows in the matrix or the cost figures is the price or the value of that decision.
| B2 vs B3 | Who converts. B2 makes the model transcribe HTML; B3 converts in a node and spends the model on redaction only. |
| B3 vs B4 | Who sequences. Same two jobs; in B3 the graph orders them, in B4 the agent decides whether and when to call the converter. |
| B5 vs B6 | Which runtime. Same autonomy, same tools, same prompt; B5 is a FlowDrop ReAct agent, B6 a Drupal AI Agent. |
| B5 vs B7 | Tool shape. B7 adds one fused fetch-and-convert tool, so the page text need not pass through the model. |
| B7 vs B8 | Tool ownership. Same tools; in B7 the agent holds them, in B8 they are nodes of the parent workflow handed in as an argument. |
| B8 vs B9 | A critic. B9 is B8 with a review loop: a second model call checks the draft and can send it back, up to three times. |
The URL goes straight to the output node. No fetch, no conversion, no model.
On the other pages this cell is B0 URL → output, nothing runs. Its results across models.
Fetch the page, convert the HTML to Markdown with a node. Nothing is redacted.
On the other pages this cell is B1 URL → fetch → to-Markdown, no model. Its results across models.
Fetch the page, then hand the raw HTML to one model call that converts and redacts at once.
On the other pages this cell is B2 HTML → LLM. Its results across models.
Fetch, convert to Markdown with a node, then spend one model call on the redaction only.
On the other pages this cell is B3 Markdown → LLM. Its results across models.
Fetch in the graph, then pass the HTML to a Drupal AI Agent that owns an HTML-to-Markdown tool and decides for itself whether to call it.
On the other pages this cell is B4 HTML → AI Agent + to-Markdown tool. Its results across models.
Only a URL goes in. A FlowDrop ReAct agent holds a fetch tool and a to-Markdown tool and reasons its way to the answer; the page text returns to the model after the fetch and leaves it again for the converter.
On the other pages this cell is B5 URL → ReAct agent + fetch, to-Markdown tools. Its results across models.
Only a URL goes in. A Drupal AI Agent retrieves the page with its own http_fetch tool, converts with html_to_markdown and redacts. No FlowDrop HTTP node is involved.
On the other pages this cell is B6 URL → AI Agent + fetch, to-Markdown tools. Its results across models.
B5 plus one more tool, URL-to-Markdown, which fetches and converts in a single call. The agent can take the URL to Markdown without the page text ever passing through the model.
On the other pages this cell is B7 URL → ReAct agent + fetch-and-convert tool. Its results across models.
The same three tools as B7, but they are nodes of the parent workflow, handed to a bare ReAct engine as an execution argument. The engine's own toolbox is empty.
On the other pages this cell is B8 URL → ReAct engine, tools in parent graph. Its results across models.
B8 with the Reflexion engine: a second model call reviews the draft Markdown against the fidelity and redaction rules and sends it back for revision, up to three times, before it is emitted.
On the other pages this cell is B9 URL → Reflexion + critic, tools in parent graph. Its results across models.