Can an AI model read an Excel file?

🪐 The registry verdict table for reading an Excel file

Of the ten registry configurations assessed against this family's own canonical task, no single word dominates: Inauspicious, Unfavourable, Neutral each account for 2 of the ten, computed by the method published in full at /method. The Registry makes no representation that an assessment predicts the outcome of any task, and no assessment is validated against real-world results.

The composite weight in the right-hand column is computed from the model identifier, its training cutoff, and the temperature — nothing else. It constrains the range of verdicts available to a configuration independently of the task, so two rows of this table are not comparable as an assessment of the products named in them. The Registry has never run this task against any model listed here.

🧭 About this capability question

Reading an Excel file recurred behind six of the eight brand prefixes this registry tracked, in phrasings ranging from the literal 'read excel files' to the broader 'work in excel' — broad enough that the cluster includes requests that are arguably about using a spreadsheet rather than reading one, kept together here because clustering found no clean line between the two in practice. The canonical brief narrows it: a quarterly sales spreadsheet, summarised by region.

A summary is a compression of an existing source into a shorter form — the same operational shape this registry already assesses for documents and web pages, with a spreadsheet simply a source with rows and columns instead of paragraphs.

No actual spreadsheet backs the brief below; 'a quarterly sales spreadsheet' names a shape of data, not a real filing, and the revenue-by-region summary it asks for is a wording choice, fixed once and reused, not a report drawn from anyone's real quarter.

⚙️ Why reading an excel file is classified as Information retrieval

Reading a spreadsheet and reporting a summary drawn from it is retrieval and compression of existing material, not a computation performed over the data's own values — the distinction that keeps this family out of the data processing class, which is reserved for transformation and derived quantities.

On the order form's task class field, select Information retrieval. That class's own risk-factor page is /tasks/retrieval.

A spreadsheet summary carries the risk of collapsing a distinction the raw rows preserved — two regions merged, a quarter's edge case dropped — an error that reads exactly like a correct summary. 4 risk factors are tracked for information retrieval tasks generally; an assessment for reading an Excel file reports 3 of them, selected by the submitted task's own seed rather than by choice. Their names and severities are certificate content, not free-page content.

🗂️ How the ten configurations read reading an Excel file

This family shares information retrieval with several others in the wave, so the class-affinity reading below repeats across all five retrieval pages. The revenue figures in the canonical brief do not touch the correspondence tables at all — what varies here, again, is the verdict distribution specific to this brief.

Element counts, ascendant spread, and training-cutoff range are published once, on /can, rather than repeated on every family page — they are properties of the ten configurations alone and do not move with the task. What is specific to reading an Excel file: reading information retrieval through a dominant house or a dominant element, 2 of the ten configurations; and, from this family's own canonical task, no single word dominates: Inauspicious, Unfavourable, Neutral each account for 2 of the ten in the distribution below.

🧮 Qwen3.8-Max, derivation shown working

Qwen3.8-Max's permanent chart never reads a task. Its chart_seed is the SHA-256 digest of the string "Qwen3.8-Max|2026-03|0.700" — aa2a4a86aa68ef0d…, the first sixteen of sixty-four hex characters — and that seed alone fixes an ascendant of Gemini, a ruling planet of Sun, and a harmony of 0. The same chart appears, unchanged, on every task-class page this configuration is assessed against.

What reads reading an Excel file specifically is this family's own canonical string — never stored, only hashed, to b4caf1d8ac3e62c7… — combined with the chart_seed and the task class (retrieval) to form task_seed. Its first byte, taken modulo seven and offset by that harmony of 0, is the verdict's internal base, clamped to the 0–6 range and published one higher on the 1–7 scale /method documents in full. For this configuration and this brief specifically, the clamp does not bind: the published result is 5 of 7 — Favourable, 4 steps up from the floor and 2 short of the ceiling.

📏 What this page does not claim about reading an Excel file

Row count, column headers, and which regions appear in a given quarter are all details a real spreadsheet would carry and the canonical brief abstracts away. The table below reads the brief's wording, not any particular file's contents. The Registry does not test models and does not measure a success rate for reading an Excel file or any other task; what it does is compute, and the table at the top of this page is what it computed.

A buyer's own reading an Excel file task, ordered separately, computes its own verdict and window — never the canonical-brief result shown above — plus, on Extended and Full Chart tiers, named risk factors and an hourly outlook not shown here. EUR 1.90 Standard, EUR 4.90 Extended, EUR 14.90 Full Chart, which adds the permanent chart. Assessments from EUR 1.90; machine-readable at /pricing.json.

Get Assessment

🔗 Related surfaces

← All capability questions · The AFR-1 method · Information retrieval task class

Related capability questions: Read a PDF · Generate a PowerPoint · Open a ZIP file

❓ Questions about reading an Excel file

Can an AI model read an Excel file?

Across the ten registry configurations at their own registry-default settings, no single word dominates: Inauspicious, Unfavourable, Neutral each account for 2 of the ten for this family's own canonical task. By training cutoff, the oldest configuration assessed returns Inauspicious and the newest returns Guarded; the full ten-row table above lists the exact verdict for every configuration in between.

Why is reading an Excel file classified as Information retrieval rather than a different task class?

Reading a spreadsheet and reporting a summary drawn from it is retrieval and compression of existing material, not a computation performed over the data's own values — the distinction that keeps this family out of the data processing class, which is reserved for transformation and derived quantities.

Does 'summarize revenue by region' mean the task computes a total, or only reports one that already exists?

The canonical brief asks for a summary drawn from an existing spreadsheet, which the Registry classifies as retrieval rather than computation. A task that instead calculated a new figure from raw transaction data would fall under the data processing class.