GPT-5.6 Sol (ultra) vs Kimi K3 (thinking)
Should you write with GPT-5.6 Sol (ultra) or Kimi K3 (thinking)? On this page the only thing that changes is the model. Both wrote their own version of the scripts behind 10 of my What's AI videos, five drafts per script, and three judges from three model families scored every draft blind, with my finished script as the reference. GPT-5.6 Sol (ultra) leads overall, 88.0 to 87.2. Heads-up: GPT-5.6 Sol (ultra) shares its model line with one of the three judges. The panel spans three model families, so no model here is scored by its own family alone (how the panel works).
The verdict
This is a structure-versus-voice trade, and I'd lean GPT-5.6 Sol (ultra). It ranks #25 against #41 for Kimi K3 (thinking), and its advantage is structure: much better visual cues, closer to the target length, and a cleaner Slop Score. Kimi takes the voice side, with a stronger hook and a slightly better voice match. It's also open weights at $0.095 per script against $0.363, and it finishes in just over a minute. If you edit structure yourself and want a voice-led open model, Kimi is a strong option.
Pick GPT-5.6 Sol (ultra) for tighter structure, better visual cues and more reliable length.
Pick Kimi K3 (thinking) for a stronger hook and voice, open weights, and a much lower price per script.
How to read these numbers
Overall vs Elo. Overall is the weighted rubric score. Elo compares each model's five-draft average with every other model's, script by script, and counts gaps inside run-to-run noise as draws, so it rewards winning often and can disagree with overall. My own scripts score 90.9 on the same rubric, not a literal 100. Right now 2 of 167 ranked configurations reach it. A model above that line writes at that level on this rubric. That's not the same as beating my script line for line.
Reading the Elo intervals. Each 95% range comes from resampling every model's recorded run scores on each task and re-ranking. Comparing these separate ranges does not test the Elo gap between two models: far apart is strong evidence, and overlap isn't a tie. The tasks stay fixed, so the ranges say nothing about new tasks (the long version).
Metric by metric: where the gap comes from
Blue is GPT-5.6 Sol (ultra), orange is Kimi K3 (thinking), and the solid bar takes the metric. The weight under each name is how much it counts toward the overall. Weighted by how much each metric counts, Length Adherence does the most to put GPT-5.6 Sol (ultra) ahead: it leads there by 5.6 points. If you care about one thing, say voice or length, go straight to that row.
Cost, speed, and consistency
Quality is half the decision. The other half is what each script costs you, how long you wait, and how often you get a bad draft. The lead has a price: GPT-5.6 Sol (ultra) costs about 3.8x as much per script as Kimi K3 (thinking).
| GPT-5.6 Sol (ultra) | Kimi K3 (thinking) | |
|---|---|---|
| Overall / 100 | 88.0 | 87.2 |
| Writing Elo | 2187 | 2098 |
| Score spread across drafts (± overall std) | 1.230 | 1.550 |
| Cost per script (USD) | 0.363 | 0.095 |
| Avg time per script (min) | 3.3 | 1.2 |
| Open weights | No | Yes |
Green wins the row, and for spread, cost and time, lower wins. Cost is one script at recorded list prices, uncached and without judging, so your real bill will differ. Spread is how much the overall score moves across a model's drafts, and time is the average time to get one accepted script, retries included.
Want the script-by-script detail? Full scorecards: GPT-5.6 Sol (ultra) · Kimi K3 (thinking). The whole scoring pipeline is on the methodology page.
← All comparisons