Can an AI model summarize a YouTube video?
The registry verdict table for summarizing a YouTube video
Of the ten registry configurations assessed against this family's own canonical task, Inauspicious for 4 of them — the modal result, 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.
| Model | Manufacturer | Training cutoff | Verdict | Score | Harmony |
|---|---|---|---|---|---|
| Gemini 3.7 Flash | Google DeepMind | 2025-01 | Inauspicious | 1 of 7 | -4 |
| DeepSeek V4 Pro | DeepSeek | 2025-12 | Neutral | 4 of 7 | -1 |
| Claude Fable 5 | Anthropic | 2026-01 | Highly auspicious | 7 of 7 | +5 |
| Claude Sonnet 5 | Anthropic | 2026-01 | Inauspicious | 1 of 7 | -3 |
| GPT-5.6 Sol | OpenAI | 2026-02 | Favourable | 5 of 7 | +1 |
| GPT-5.6 Terra | OpenAI | 2026-02 | Auspicious | 6 of 7 | -1 |
| Grok 4.6 | xAI | 2026-02 | Inauspicious | 1 of 7 | -1 |
| Kimi K3 | Moonshot AI | 2026-03 | Unfavourable | 2 of 7 | -1 |
| Qwen3.8-Max | Alibaba Cloud | 2026-03 | Inauspicious | 1 of 7 | 0 |
| Claude Opus 5 | Anthropic | 2026-05 | Highly auspicious | 7 of 7 | +2 |
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
Summarizing a YouTube video recurred behind six of the eight brand prefixes tracked, in a cluster kept deliberately narrow: 'summarize youtube videos' and its close relatives, held apart from the separate 'watch youtube videos' request, which describes passive consumption rather than a deliverable. The canonical brief is the summarizing shape — a video watched, and a written account of its key points produced.
'Watch' is the operative verb in that brief, and it names the same retrieval shape this wave's other summarization and extraction families share: something that already exists is consulted, and a shorter written account of it is produced.
No actual video sits behind the canonical brief below. It fixes a task shape — a video watched, key points summarized — reused unchanged on every load, so the verdict computed from it describes that shape and not the content of any real recording.
How the ten configurations read summarizing a YouTube video
This is the fifth and last of the wave's retrieval-class families, so the dominant-house and dominant-element reading below matches what its sibling pages already show. The verdict distribution below is where this canonical brief's own computed answer lives.
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 summarizing a YouTube video: reading information retrieval through a dominant house or a dominant element, 2 of the ten configurations; and, from this family's own canonical task, Inauspicious for 4 of them — the modal result in the distribution below.
| Published value | Verdict | Configurations |
|---|---|---|
| 1 of 7 | Inauspicious | 4 of the ten |
| 2 of 7 | Unfavourable | 1 of the ten |
| 3 of 7 | Guarded | 0 of the ten |
| 4 of 7 | Neutral | 1 of the ten |
| 5 of 7 | Favourable | 1 of the ten |
| 6 of 7 | Auspicious | 1 of the ten |
| 7 of 7 | Highly auspicious | 2 of the ten |
Claude Opus 5, derivation shown working
Claude Opus 5's permanent chart never reads a task. Its chart_seed is the
SHA-256 digest of the string "Claude Opus 5|2026-05|0.700"
— 4b294997ecca0cfe…, the first sixteen of sixty-four hex
characters — and that seed alone fixes an ascendant of Cancer, a
ruling planet of Saturn, and a harmony of +2.
The same chart appears, unchanged, on every task-class page this configuration is
assessed against.
What reads summarizing a YouTube video specifically is this family's own canonical
string — never stored, only hashed, to c430d394cdeed557… —
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
+2, 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.
Nothing left to clamp toward here: for this brief, this configuration's internal score sits at the ceiling, published as 7 of 7 — Highly auspicious.
Why summarizing a youtube video is classified as Information retrieval
Producing a written summary of an existing video's content is retrieval and compression, the same operational category applied to summarizing a document or a web page — the medium changes, the operation does not.
On the order form's task class field, select Information retrieval. That class's own risk-factor page is /tasks/retrieval.
A video summary carries the risk of compressing away the one point a downstream reader actually needed — technically accurate about most of the video, silent about the part that mattered. 4 risk factors are tracked for information retrieval tasks generally; an assessment for summarizing a YouTube video 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.
What this page does not claim about summarizing a YouTube video
Video length, spoken language, and whether captions exist are all properties a real video would carry and the canonical brief does not specify. The table below reads the brief's wording, not any particular recording. The Registry does not test models and does not measure a success rate for summarizing a YouTube video 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 summarizing a YouTube video 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.
Related surfaces
← All capability questions · The AFR-1 method · Information retrieval task class
Related capability questions: Transcribe audio · Read a PDF · Open a link
Questions about summarizing a YouTube video
- Can an AI model summarize a YouTube video?
- Across the ten registry configurations at their own registry-default settings, Inauspicious for 4 of them — the modal result for this family's own canonical task. By training cutoff, the oldest configuration assessed returns Inauspicious and the newest returns Highly auspicious; the full ten-row table above lists the exact verdict for every configuration in between.
- Why is summarizing a YouTube video classified as Information retrieval rather than a different task class?
- Producing a written summary of an existing video's content is retrieval and compression, the same operational category applied to summarizing a document or a web page — the medium changes, the operation does not.
- Does the method watch or transcribe the video before computing a verdict?
- No. The method hashes the wording of the canonical brief published above; no video has been watched, transcribed, or otherwise processed in producing the table on this page.