Benchmark
LiveCodeBench v6
What is LiveCodeBench v6?
LiveCodeBench is a holistic and contamination-free evaluation benchmark for large language models for code. It continuously collects new problems from…
- Modality
- text
- Evaluates
- reasoning, general
- Openness
- unknown
- Importer
- LLM Stats
- Reported scores
- 56
Source provenance
LiveCodeBench v6 paper and code
LiveCodeBench v6 results and reported scores
LLM Stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| DiffusionGemma 26B-A4B | 0.691 | 2026-06-10 | evidence | |
| Gemma 4 12B | 0.72 | 2026-05-23 | evidence | |
| Gemma 4 26B-A4B | 0.771 | 2026-04-02 | evidence | |
| Gemma 4 31B | 0.8 | 2026-04-02 | evidence | |
| Gemma 4 E2B | 0.44 | 2026-04-02 | evidence | |
| Gemma 4 E4B | 0.52 | 2026-04-02 | evidence | |
| GLM-4.6 | Z.ai | 0.828 | 2025-09-30 | evidence |
| GLM-4.7 | Z.ai | 0.849 | 2025-12-22 | evidence |
| GPT OSS 120B High | OpenAI | 0.819 | 2025-08-05 | evidence |
| K-EXAONE-236B-A23B | LG AI Research | 0.807 | 2025-12-31 | evidence |
| Kimi K2 Base | Moonshot AI | 0.263 | 2025-07-11 | evidence |
| Kimi K2 Instruct | Moonshot AI | 0.537 | 2025-07-11 | evidence |
| Kimi K2-Thinking-0905 | Moonshot AI | 0.831 | 2025-09-05 | evidence |
| Kimi K2.5 | Moonshot AI | 0.85 | 2026-01-27 | evidence |
| Kimi K2.6 | Moonshot AI | 0.896 | 2026-04-20 | evidence |
| LFM2.5-2.6B | Liquid AI | 0.5941 | 2026-08-04 | evidence |
| MAI-Thinking-1 | Microsoft | 0.877 | 2026-06-02 | evidence |
| MiMo-V2-Flash | Xiaomi | 0.806 | 2025-12-16 | evidence |
| MiMo-V2.5-Pro | Xiaomi | 0.396 | 2026-04-27 | evidence |
| MiniCPM-SALA | OpenBMB | 0.52 | 2026-02-11 | evidence |
| Nemotron 3 Nano (30B A3B) | NVIDIA | 0.683 | 2025-12-15 | evidence |
| Nemotron 3 Ultra (550B A55B) | NVIDIA | 0.89 | 2026-06-04 | evidence |
| North Mini Code 1.0 | Cohere | 0.703 | 2026-06-09 | evidence |
| Qwen3 Max | Qwen | 0.69 | 2025-12-15 | evidence |
| Qwen3 VL 235B A22B Instruct | Qwen | 0.543 | 2025-09-22 | evidence |
| Qwen3 VL 235B A22B Thinking | Qwen | 0.701 | 2025-09-22 | evidence |
| Qwen3 VL 30B A3B Instruct | Qwen | 0.426 | 2025-09-22 | evidence |
| Qwen3 VL 30B A3B Thinking | Qwen | 0.642 | 2025-09-22 | evidence |
| Qwen3 VL 32B Instruct | Qwen | 0.438 | 2025-09-22 | evidence |
| Qwen3 VL 32B Thinking | Qwen | 0.656 | 2025-09-22 | evidence |
| Qwen3 VL 4B Instruct | Qwen | 0.379 | 2025-09-22 | evidence |
| Qwen3 VL 4B Thinking | Qwen | 0.513 | 2025-09-22 | evidence |
| Qwen3 VL 8B Instruct | Qwen | 0.393 | 2025-09-22 | evidence |
| Qwen3 VL 8B Thinking | Qwen | 0.586 | 2025-09-22 | evidence |
| Qwen3-235B-A22B-Instruct-2507 | Qwen | 0.518 | 2025-07-22 | evidence |
| Qwen3-235B-A22B-Thinking-2507 | Qwen | 0.741 | 2025-07-25 | evidence |
| Qwen3-Next-80B-A3B-Instruct | Qwen | 0.566 | 2025-09-10 | evidence |
| Qwen3-Next-80B-A3B-Thinking | Qwen | 0.687 | 2025-09-10 | evidence |
| Qwen3.5-122B-A10B | Qwen | 0.789 | 2026-02-24 | evidence |
| Qwen3.5-27B | Qwen | 0.807 | 2026-02-24 | evidence |
| Qwen3.5-35B-A3B | Qwen | 0.746 | 2026-02-24 | evidence |
| Qwen3.5-397B-A17B | Qwen | 0.836 | 2026-02-16 | evidence |
| Qwen3.5-4B | Qwen | 0.558 | 2026-03-02 | evidence |
| Qwen3.5-9B | Qwen | 0.656 | 2026-03-02 | evidence |
| Qwen3.6 Plus | Qwen | 0.871 | 2026-04-02 | evidence |
| Qwen3.6-27B | Qwen | 0.839 | 2026-04-21 | evidence |
| Qwen3.6-35B-A3B | Qwen | 0.804 | 2026-04-16 | evidence |
| Qwen3.7 Max | Qwen | 0.916 | 2026-05-19 | evidence |
| Qwen3.7-Plus | Qwen | 0.896 | 2026-05-31 | evidence |
| Qwen3.8-27B | Qwen | 0.903 | 2026-08-14 | evidence |
| Sakana Namazu | Sakana AI | 0.9033 | 2026-08-03 | evidence |
| Sarvam-105B | Sarvam AI | 0.717 | 2026-03-06 | evidence |
| Sarvam-30B | Sarvam AI | 0.7 | 2026-03-06 | evidence |
| Seed 2.0 Lite | ByteDance | 0.817 | 2026-02-14 | evidence |
| Seed 2.0 Pro | ByteDance | 0.878 | 2026-02-14 | evidence |
| Step-3.5-Flash | StepFun | 0.864 | 2026-02-02 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.