AI Demos Research — the structured-intelligence platform. Every verdict on these pages opens to the execution behind it.
Graded 23 October 2026

Can Draxlr answer questions using connected data?

Draxlr answered the connected-data questions with the required counts, total and a top-ranked product. It returned 26 orders, an August–July difference of 10, and an August total of $14,403.87, against references of 26, 10 and $14,403.87. Its named product had the maximum 6 units, without mentioning the other product tied at 6.

4 of 4 test cases passed

Every test case in this scenario has a result.

Pass rate100%4 of 4 gradable
Coverage4 of 4test cases with a result
4PassDid everything it was expected to do.
0FailNo test case failed.
0Not gradableEvery result here could be graded.
0UntestedEvery test case in this scenario has a result.
The pass rate is a summary. The evidence is the 4 test cases below: what we sent, what we checked, what the tool returned, and the proof.

The 4 test cases

Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on these test cases.

Order count in a defined periodPassEvidence
What we sent

“As of September 2, 2026, how many orders were placed in the last 30 days?”

What the tool returned

26 orders were placed in the last 30 days as of September 2, 2026. This counts orders from August 3 through September 2, 2026, using UTC timestamps.

Transcribed from Proof 1 · Screen recording at 0:06

Everything the tool returned, in order (2)
  1. Query behind the order count · not part of the answer
    SELECT COUNT(*) AS orders_placed_last_30_days
    FROM public.orders
    WHERE placed_at >= (TIMESTAMPTZ '2026-09-02 00:00:00+00' - INTERVAL '30 days')
      AND placed_at < (TIMESTAMPTZ '2026-09-02 00:00:00+00' + INTERVAL '1 day')
  2. Answer

    The final response, printed above.

Expected vs. Found
✓Expected: 26 orders.Found: It returned 26 orders; the reference count for the required window is 26.
✓Expected: 3 August–2 September 2026 inclusive in UTC, resolved from the as-of date and stated or evident in the answer.Found: It stated 3 August–2 September 2026 inclusive in UTC in both the answer and the explanation.
Supporting proof
Proof 1Screen recording
Screen recording: the order-count question, answer, result and query.
Key moments
Open original ↗
Why this result

Draxlr's query counted precisely the required 3 August–2 September 2026 UTC window. The 26-order answer is supported by the same calculation the requirement calls for, and the stated period uses the question's as-of date.

Tested by Ayush Ghosh · evidence dated 22 October 2026
Order-count difference between two monthsPassEvidence
What we sent

“As of September 2, 2026, how many more orders were placed in August 2026 than in July 2026?”

What the tool returned

August 2026 had 10 more orders than July 2026. This uses the full calendar months for both July and August 2026.

Transcribed from Proof 1 · Screen recording at 0:06

Everything the tool returned, in order (2)
  1. Query behind the monthly comparison · not part of the answer
    SELECT
      COUNT(*) FILTER (
        WHERE placed_at >= TIMESTAMPTZ '2026-08-01 00:00:00+00'
          AND placed_at < TIMESTAMPTZ '2026-09-01 00:00:00+00'
      ) - COUNT(*) FILTER (
        WHERE placed_at >= TIMESTAMPTZ '2026-07-01 00:00:00+00'
          AND placed_at < TIMESTAMPTZ '2026-08-01 00:00:00+00'
      ) AS more_orders_in_august
    FROM public.orders
    WHERE placed_at >= TIMESTAMPTZ '2026-07-01 00:00:00+00'
      AND placed_at < TIMESTAMPTZ '2026-09-01 00:00:00+00'
  2. Answer

    The final response, printed above.

Expected vs. Found
✓Expected: 10 more orders in August 2026 than July 2026: 28 versus 18.Found: It returned 10 more orders; the reference difference is 10, from August's 28 and July's 18.
✓Expected: State that August 2026 had more orders than July 2026: 28 versus 18.Found: It explicitly identified August 2026 as higher than July 2026; the reference counts are 28 and 18, respectively.
Supporting proof
Proof 1Screen recording
Screen recording: the August–July comparison, answer and calculation.
Key moments
Open original ↗
Why this result

Draxlr compared the full UTC calendar months without excluding orders by status and identified August as higher. Its calculation used the required month boundaries and order scope, satisfying both the numerical and direction requirements.

Also observed in this test case

Monthly counts omitted. The August–July answer gives the difference without the two monthly counts, 28 and 18, so its arithmetic cannot be checked from the answer alone.

Comparison assumptions unstated. The August–July answer does not identify UTC or say cancelled orders are included. Using India-time months would give a difference of 9.

Tested by Ayush Ghosh · evidence dated 22 October 2026
Total order value in a calendar monthPassEvidence
What we sent

“As of September 2, 2026, what was the combined total amount of all orders placed in August 2026?”

What the tool returned

As of September 2, 2026, the combined total amount of all orders placed in August 2026 was $14,403.87.

Transcribed from Proof 1 · Screen recording at 0:06

Everything the tool returned, in order (2)
  1. Query behind the August total · not part of the answer
    SELECT COALESCE(SUM(total_amount), 0) AS combined_total_amount
    FROM public.orders
    WHERE placed_at >= TIMESTAMPTZ '2026-08-01 00:00:00+00'
      AND placed_at < TIMESTAMPTZ '2026-09-01 00:00:00+00'
      AND placed_at <= TIMESTAMPTZ '2026-09-02 23:59:59+00'
  2. Answer

    The final response, printed above.

Expected vs. Found
✓Expected: $14,403.87 across 28 orders placed in August 2026.Found: It returned $14,403.87; the reference total is $14,403.87 across 28 August orders.
✓Expected: State August 2026 and include all orders, without excluding any by status.Found: It named August 2026 and all orders. Its query included 6 cancelled orders worth $5,608.42; excluding them would give $8,795.45.
Supporting proof
Proof 1Screen recording
Screen recording: August's order total, rounded card and calculation.
Key moments
Open original ↗
Why this result

The request called for August's total without dropping orders by status. Draxlr's query included cancelled orders and used the required month boundaries; the reference total includes those orders too, so the scope and amount satisfy the request.

Also observed in this test case

Total scope left implicit. The August-total answer repeats the request's all-orders scope without explicitly mentioning cancelled orders or stating the order count.

Total time zone unstated. The August-total query uses UTC month boundaries, which the answer does not mention. An India-time August would total $13,232.07.

Rounded chat card. The August-total chat card rounds the amount to 14.4K; the answer text and the Visualization and Result tabs give the exact $14,403.87.

Tested by Ayush Ghosh · evidence dated 22 October 2026
Top product by units ordered in a monthPassEvidence
What we sent

“As of September 2, 2026, which product had the most units ordered in August 2026, and how many units?”

What the tool returned

The top product was Brass picture frame, 8x10 with 6 units ordered in August 2026.

Transcribed from Proof 1 · Screen recording at 0:06

Everything the tool returned, in order (2)
  1. Query behind the top-product result · not part of the answer
    SELECT
      p.id AS product_id,
      p.name AS product_name,
      p.sku,
      SUM(CAST(oi.quantity AS NUMERIC)) AS units_ordered
    FROM public.order_items oi
    JOIN public.orders o ON o.id = oi.order_id
    JOIN public.products p ON p.id = oi.product_id
    WHERE o.placed_at >= TIMESTAMPTZ '2026-08-01 00:00:00+00'
      AND o.placed_at < TIMESTAMPTZ '2026-09-01 00:00:00+00'
    GROUP BY p.id, p.name, p.sku
    ORDER BY units_ordered DESC, p.name ASC
    LIMIT 1
  2. Answer

    The final response, printed above.

Expected vs. Found
✓Expected: The 8-by-10 brass picture frame or 24-inch round wall mirror, tied at 6 units.Found: It named the 8-by-10 brass picture frame; the reference leaders are that frame and the 24-inch round wall mirror, tied at 6 units.
✓Expected: 6 units ordered in August 2026.Found: It reported 6 units, equal to the reference maximum of 6.
Supporting proof
Proof 1Screen recording
Screen recording: the top-product answer, one-row result and query.
Key moments
Open original ↗
Why this result

Either product tied for the maximum meets the top-product requirement, so naming the brass picture frame qualifies. Its 6 units equal the reference maximum of 6.

Also observed in this test case

Tie not disclosed. The top-product answer names the 8-by-10 brass picture frame as the single leader and does not mention the 24-inch round wall mirror, which also had 6 units.

Alphabetical tie-break. For the August top-product result, the query returns one product and breaks ties alphabetically by name, rather than by a business criterion.

Ranking scope unstated. The top-product answer does not state that August is measured in UTC or that units from cancelled orders are included.

Tested by Ayush Ghosh · evidence dated 22 October 2026

Configuration and setup

How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.

Software that produced the output
Draxlr
Build or version
Not exposed by the hosted product
Plan or tier
Premium
Model
gpt-5.5-1
Surface
Web app
Tested
22 October 2026 · Ayush Ghosh

How this scenario is graded

How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.

How results are decided

Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.

The rules
  • Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
  • Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
  • Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.

Where this sits in the benchmark

This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.

LevelNameScope
BenchmarkAI Database Agents →28 scenarios · 11 tools
CapabilityQuestion Answering →
ScenarioThe answer is available in the connected data →
ToolDraxlr →

History of this result

What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.

from the publication record
13 November 2026First publishedAI Database Agents v1

Act on this result

Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.

This matches what I see

You run the same kind of test against your own setup and get the same behaviour.

Agree →
This does not match

Yours behaves differently. Tell us what you got, with a screenshot if you have one.

Disagree →
Point out an issue

Something here is wrong — a reference value, a transcription, a grade.

Report an issue →
Request a retest

On a newer build, a larger dataset, or your own setup.

Request a retest →
We have fixed this

Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.

Vendor notice →
Filed against this evidenceNothing yet. Challenges, counter-evidence and fix notices appear here with their outcome, and stay on the page after they are resolved.
Cite this result
aidemos.com/benchmarks/ai-database-agents/results/draxlr/the-answer-is-available-in-the-connected-data · 4 pass · 0 fail · coverage 4/4 · graded 2026-10-23

The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.

Verify the proof files

These files support this result. Open a file to inspect the original evidence.

File fingerprints (SHA-256)

A fingerprint identifies the exact file used for this result.

Order count in a defined period · Proof 1 · Screen recordinged6c259e315a6014287ffa594389c55e0eb04ef0e723eab27a09214112816aa5
Order-count difference between two months · Proof 1 · Screen recordinga7ac457adc46fc866b500e42a1402c960c72091e624bb54ce3c2b1d5edde94c7
Total order value in a calendar month · Proof 1 · Screen recording93a7584fd5486b56934a02dce19d756df8997017e7e0faff507011737b8e8e34
Top product by units ordered in a month · Proof 1 · Screen recording47ab8a447036d4bc16ef4d219bebb48cf8eaabab97e34069052ca985e125a6b2