Media generation: Task risk assessment
An agent operating in this task class produces an artefact that did not exist before the task began — an image, a video clip, a synthesised voice, a music stem, a slide deck, or any other generated file assembled against a stated brief. Nothing is retrieved and nothing is transformed. The output is manufactured from a specification, and that specification is the only fixed point the result can ever be checked against. Across images, video, audio, speech, and slides the operational shape repeats without exception: a brief goes in, an artefact comes out, and the distance between the two is measured in words rather than in a diff.
Verification runs backwards here relative to the other six classes in this registry. A retrieval task can be checked against the source it drew from, and a data processing task against the number it should have produced; a generated artefact has no equivalent to check against, only the brief it was asked to satisfy, read after delivery by whoever approves it. That reading is not something this registry performs. What the registry computes is a risk profile for the task and a permanent chart for the model submitted against it — not a rating of the artefact, and not an opinion on whether the image, the clip, or the deck turned out well.
What a media generation assessment computes
For media generation tasks, the task-level portion of a full assessment draws three of the four risk factors above and pairs them with the model's own permanent profile — ascendant, ruling planet, element, and aspects — computed once per configuration and not per artefact. What the method does not do is open the file: no image is viewed, no audio is played, no rendered slide is reviewed, because the assessment is computed rather than inspected. The complete derivation is documented in the open at /method.
Four ways a generated artefact fails
Four risk factors are tracked for media generation, and their names describe what happens to an artefact after it is produced rather than a flaw in the brief that produced it. All four are tracked on every media generation assessment the registry issues; three are reported on any one of them, and which three is derived from the task rather than chosen by whoever placed the order.
The Registry tracks 4 risk factors for media generation tasks. They are fixed for the class: they do not adapt to a model, to a configuration, or to the wording of a task. An assessment reports 3 of the 4, each with its severity rating and its description, and which 3 are returned is derived from the submitted task's own seed rather than chosen. The factors themselves are part of the assessment and are not published here.
Media generation crossed against every registry model
Media generation crossed against each registry model. As on every other task class page in this registry, nothing in a profile is drawn from a prompt, a resolution, or a render duration — a profile is computed once from a model's own configuration and carries over onto this page unchanged.
The ordering follows the same rule as every crossing table in this registry: 0 of the 10 profiles read media generation through a dominant house or an element, 0 of them through the house alone. Claude Fable 5 leads at 0 and Qwen3.8-Max closes the table with a profile that points elsewhere. The 10 models still resolve to 6 distinct ruling planets, so the ruling-planet column is not one entry repeated down the page.
Every row is computed from that model's registry default parameters — the training cutoff listed beside it, at temperature 0.700. A media generation certificate is computed from the exact parameters submitted with the order instead.
| Model | Manufacturer | Training cutoff | Ruling planet | Reads this class through | Outlook |
|---|---|---|---|---|---|
| Claude Fable 5 | Anthropic | 2026-01 | Mars | neither — class-neutral (0) | publication pending |
| Claude Opus 5 | Anthropic | 2026-05 | Saturn | neither — class-neutral (0) | publication pending |
| Claude Sonnet 5 | Anthropic | 2026-01 | Venus | neither — class-neutral (0) | publication pending |
| GPT-5.6 Sol | OpenAI | 2026-02 | Mercury | neither — class-neutral (0) | publication pending |
| GPT-5.6 Terra | OpenAI | 2026-02 | Moon | neither — class-neutral (0) | publication pending |
| Grok 4.6 | xAI | 2026-02 | Sun | neither — class-neutral (0) | publication pending |
| Gemini 3.7 Flash | Google DeepMind | 2025-01 | Venus | neither — class-neutral (0) | publication pending |
| Kimi K3 | Moonshot AI | 2026-03 | Moon | neither — class-neutral (0) | publication pending |
| DeepSeek V4 Pro | DeepSeek | 2025-12 | Sun | neither — class-neutral (0) | publication pending |
| Qwen3.8-Max | Alibaba Cloud | 2026-03 | Sun | neither — class-neutral (0) | publication pending |
Questions about media generation assessment
- What counts as a media generation task for this registry?
- Producing an image, a video clip, an audio or speech file, a slide deck, or any other newly generated artefact falls under this task class, whatever the generation tool and whatever the output format.
- How many risk factors does an assessment report?
- The registry tracks 4 risk factors for this class and an assessment reports 3 of them, each with its severity rating and its description. Which 3 are returned is derived from the submitted task's own seed. The factors are part of the assessment and are not published on this page.
- Does a media generation assessment evaluate the quality of the image, video, or audio produced?
- No. The registry computes a risk profile and a permanent chart from the submitted configuration; it does not open, play, or render the artefact, and it makes no claim about composition, fidelity, or artistic quality.
Order a media generation assessment
A Standard Assessment costs EUR 1.90, an Extended Assessment EUR 4.90, and a Full Chart Assessment EUR 14.90 — the last of which carries the permanent chart itself. Assessments from EUR 1.90. Machine-readable pricing is published at /pricing.json, with no authentication required and no difference from the figures printed here.
The task class is selected on the order form. The verdict and the recommended execution window are computed from that selection together with the submitted parameters, and they appear on the certificate — not on this page, and not anywhere else in the directory.