LLM intelligence hub
EvalKit brings public leaderboards, model profiles, benchmark guides, and curated LLM news into one citation-first workspace.
Current leaders
Reasoning leader: 72.51. leads on reasoning
LLM StatsGoogleCheapest top 10: $0.5/M. lowest input price
LLM StatsxAILongest context: 2.0M tokens. largest context window
LLM StatsAnthropicCoding leader: 57.83. wins on coding
LLM StatsMistral AIFastest output: 678.15 c/s. highest throughput
LLM StatsMoonshot AIBest open weights: 58.13. top open-weight score
LLM StatsLLM news
Evidence synthesis is crucial for turning primary research into reliable knowledge for science, medicine, education, and policy. Yet, quantitative evidence synthesis remains largely manual and difficult to scale. Here, we introduce AutoSyn...
Security-agent evaluations commonly measure peak offensive capability under generous inference budgets, emphasizing vulnerability discovery, exploit development, penetration testing, and CTF completion. Such measurements are useful but inc...
Healthcare multimodal AI must combine visual and textual evidence while remaining reliable and interpretable. Using MediaEval Medico 2025 as a retrospective GI endoscopy case study, we analyze design choices across nine documented systems...
Retrieval systems are trained and evaluated on a static idea of usefulness: hand a document and a question to a reader model, see whether the answer improves, and score the document accordingly. The idea holds up when a document is read on...
Benchmark guide
Trust policy
Public rows are labeled as replicated public-source data or editorial context. “Verified by EvalKit” stays at 0 until there is real run evidence attached.
