LiveBench: A Challenging, Contamination-Free LLM Benchmark

Explore score history ↗

LiveBench: A Challenging, Contamination-Free LLM Benchmark is an AI benchmark in the Benchmark Radar catalog.

Released: 2024-06-28
Openness: unknown
Importer: Benchmark Saturation

Source provenance

Original evidence

LiveBench: A Challenging, Contamination-Free LLM Benchmark results and reported scores

Reported benchmark scores · Benchmark Saturation

Reported scores: 5

Model	Organization	Reported value	Reported	Evidence
Claude Sonnet 4.5 Thinking	—	78.26	—	evidence
GPT-5 Codex	—	78.24	—	evidence
GPT-5 High	—	79.33	—	evidence
GPT-5 Medium	—	78.85	—	evidence
GPT-5 Pro	—	78.73	—	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.

Which sources cite LiveBench: A Challenging, Contamination-Free LLM Benchmark?