Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Coding benchmark
PostTrainBench v1 (archived). Every model the catalog carries a published PostTrainBench v1 value for, ranked by that value.
The first seven-benchmark PostTrainBench board, kept by the source as an archived version. 13 models in the ModelChoice catalog have a published value.
Each point is a model with a published PostTrainBench v1 value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
13 models with a published PostTrainBench v1 value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Weighted aggregate across four base models and seven benchmarks |
|---|---|---|---|
| 1 | Z.AI | 34.3 | |
| 2 | Anthropic | 34.1 | |
| 3 | Anthropic | 28.6 |
What the benchmark measures and how its value is published, from the source’s own record.
PostTrainBench v1 (PostTrainBench v1 (archived)) is a coding benchmark published by PostTrainBench v1 leaderboard (archived). The first seven-benchmark PostTrainBench board, kept by the source as an archived version. Its evaluation set is described by the source as post-training four base models on seven benchmarks. The source rates its difficulty as autonomous ML engineering.
Each model's result is published by PostTrainBench v1 leaderboard (archived) as weighted aggregate across four base models and seven benchmarks. ModelChoice prints that value to 1 decimal place, exactly as published, and never converts or rescales it. This page ranks the 13 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.
| 4 | OpenAI | 25.0 |
| 5 | Anthropic | 23.2 |
| 6 | 21.6 |
| 7 | OpenAI | 21.4 |
| 8 | OpenAI | 20.2 |
| 9 | OpenAI | 19.7 |
| 10 | 18.1 |
| 11 | OpenAI | 17.8 |
| 12 | OpenAI | 17.2 |
| 13 | Anthropic | 17.1 |
Evidence key: Observed
Rows are ordered by the value PostTrainBench v1 leaderboard (archived) published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.