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
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code. Every model the catalog carries a published LiveCodeBench value for, ranked by that value.
A continuously updated coding benchmark built from newly collected LeetCode, AtCoder, and Codeforces problems. Fresh problem windows reduce one contamination path, but results still need a release and setup check. 20 models in the ModelChoice catalog have a published value.
Each point is a model with a published LiveCodeBench value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
20 models with a published LiveCodeBench 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 | Competitive-programming evaluation |
|---|---|---|---|
| 1 | Alibaba | 91.6 | |
| 2 | Alibaba | 89.6 | |
| 3 | Upstage | 87.8 |
What the benchmark measures and how its value is published, from the source’s own record.
LiveCodeBench (LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code) is a coding benchmark published by LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code. A continuously updated coding benchmark built from newly collected LeetCode, AtCoder, and Codeforces problems. Fresh problem windows reduce one contamination path, but results still need a release and setup check. Its evaluation set is described by the source as continuously updated contest problems. The source rates its difficulty as competitive programming level.
Each model's result is published by LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code as competitive-programming evaluation. ModelChoice prints that value exactly as published, and never converts or rescales it. This page ranks the 20 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 | Z.AI | 84.9 |
| 5 | Alibaba | 83.9 |
| 6 | Alibaba | 80.4 |
| 7 | DeepSeek | 74.8 |
| 7 | xAI | 74.8 |
| 9 | DeepSeek | 73.3 |
| 10 | NVIDIA | 66.31 |
| 11 | MiniMax | 65 |
| 12 | DeepSeek | 60.5 |
| 13 | DeepSeek | 57.5 |
| 14 | xAI | 57 |
| 15 | DeepSeek | 56.4 |
| 16 | Microsoft | 53.8 |
| 17 | DeepSeek | 53.1 |
| 17 | Microsoft | 53.1 |
| 19 | Alibaba | 38.7 |
| 20 | DeepSeek | 37.6 |
Evidence key: ObservedLast good
Rows are ordered by the value LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code 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.