BlogLLM Benchmark Scores as Published, Side by Side
LiveCodeBench Pro: Results as Published
Every result the catalog holds for LiveCodeBench Pro: 3 models from 2 developers, read from 1 page. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area Code GenerationMaintainer LiveCodeBench Pro team (Zihan Zheng and co-authors)Introduced 2025Data version 2026103001, as of 2026-10-30
LiveCodeBench Pro tests competitive programming with continuously updated problems from Codeforces, ICPC and the IOI, so the problems postdate the training data of the models. Olympiad medalists label the algorithmic category of every problem and review failed submissions. Publishers print either pass@1 on a difficulty tier or an Elo-style rating that places a model on the Codeforces scale.
LiveCodeBench Pro (Elo)
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-5.2OpenAI, released 2025-12-11 | 2393 | Google, Gemini 3.1 Pro model cardchecked 2026-10-15 | Effort: xhigh |
| Gemini 3.1 Pro PreviewGoogle, released 2026-02-19 | 2887 | Google, Gemini 3.1 Pro model cardchecked 2026-10-15 | Effort: high |
| Gemini 3 ProGoogle, released 2025-11-18 | 2439 | Google, Gemini 3.1 Pro model cardchecked 2026-10-15 | Effort: high |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.