Kimi K3 vs DeepSeek V4 Pro 0813 (max)
Should you write with Kimi K3 or DeepSeek V4 Pro 0813 (max)? 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. Kimi K3 leads overall, 88.5 to 84.8.
The verdict
My pick is Kimi K3. It's #21 against #56, about 300 Elo higher, and it takes all nine metrics. Continuity & Emotion is the widest split, about five and a half points, and that's the metric for whether each paragraph pulls you into the next. DeepSeek V4 Pro 0813 (max) is also the less predictable of the two: its overall score swings about three times as much from run to run. Both are open weights, so that isn't the tiebreaker. The case for DeepSeek is cost, about $0.021 a script against $0.260, more than ten times cheaper, and it gets stronger if you generate several drafts and keep the best one.
Pick Kimi K3 when you want consistent, well-connected scripts from an open-weights model.
Pick DeepSeek V4 Pro 0813 (max) when cost per script matters more than consistency and you can pick the best of several runs.
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 Kimi K3, orange is DeepSeek V4 Pro 0813 (max), 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, Continuity & Emotion does the most to put Kimi K3 ahead: it leads there by 5.5 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: Kimi K3 costs about 12x as much per script as DeepSeek V4 Pro 0813 (max).
| Kimi K3 | DeepSeek V4 Pro 0813 (max) | |
|---|---|---|
| Overall / 100 | 88.5 | 84.8 |
| Writing Elo | 2217 | 1917 |
| Score spread across drafts (± overall std) | 1.260 | 4.040 |
| Cost per script (USD) | 0.260 | 0.021 |
| Avg time per script (min) | 3.9 | 3.2 |
| Open weights | Yes | 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: Kimi K3 · DeepSeek V4 Pro 0813 (max). The whole scoring pipeline is on the methodology page.
← All comparisons