Agent routing or picking the model yourself
Three job types where an agent's model choice predictably goes wrong, and how to pin a named model in the brief without giving up routing everywhere else.
Three job types where an agent's model choice predictably goes wrong, and how to pin a named model in the brief without giving up routing everywhere else.
Vendor number-one claims name whichever board flatters them. A two-minute check on the board, the date, the vote count and the exact category behind the claim.
Finding which model holds a brief is billable work most studios absorb. How to scope a model-selection pass, cap it on three axes, and invoice it defensibly.
For batch work, how many generations you hold in flight decides wall-clock time more than per-clip speed. How to measure throughput and order a queue.
An elimination protocol that takes 296 models to 8 in five ordered gates: input modality, duration ceiling, audio, licence, credit band. Refill quarterly.
Pours, flame and volumetric smoke fail in three different ways. How to rank them, and a rule for what to generate, what to composite, and what to cut.
Single-image, three-slot and nine-slot reference models compared: which shots improve past three references, which just get slower, and a shot-type table.
General-purpose AI providers have published training-content summaries since August 2025. Six questions that separate a useful one from a decorative one.
Fast voice models and expressive ones fail in different places. A two-axis method for picking TTS per job, plus the retake maths that settles it.
Complaints that a model ignored the prompt are usually a casting error. Two prompts at three seeds each sort image models into the two camps that matter.
A reading method for the quality-per-credit frontier that shortlists models for one specific job and still works after the next round of launches.
Knowing one model deeply beats a marginally better one you cannot steer. Two thresholds, job variety and output volume, for when that stops being true.
Six spec numbers that promise more than they deliver: max duration, resolution, reference count, fps, native audio, multi-shot, and one test to run on each.
Some models reach 2K by regenerating a low-res render; others generate high-res natively. What each does to fine detail, temporal stability and render time.
When a cheap generation plus a dedicated upscaler beats the top tier, and when the upscaler invents detail you did not want. A rule keyed to four shot types.
A 15-second ceiling is rarely 15 usable seconds. The three ways long clips fall apart, a ladder test to find your safe duration, and where to place the cut.
Bowed lines, disagreeing vanishing points and drifting facade counts sink generated architecture. The traits that help, and the prompt that fixes interiors.
Background people break more shots than hero subjects. The pixel budget that predicts failure, plus staging tactics for street, stadium and venue shots.
71 text-to-image, 42 edit-image, 30 image-to-image models — and only 3 do all three. A job-first way to pick the right one instead of shopping by brand.
The sticker price of an AI video model isn't the real cost per finished shot. How to measure your own reroll rate and budget credits against it.