Public admission truth
Benchmark health
Catalog discovery, scraper implementation, admission, and rank eligibility are separate stages. This page reports each without turning missing publication into a claim that a benchmark is dead.
Benchmark coverage
See which capability cells have a published ranking, which are still in preview, and which have no implemented feed yet.
0
0 active
7
7 preview
21
21 unavailable
Health generated Jul 10, 2026, 12:00 AM UTC
Code generation
PreviewProduce correct code from a spec or prompt
2 implemented · 0 admitted · 0 rank-eligible
Autonomous SWE agent
PreviewComplete repository-level software engineering tasks
1 implemented · 0 admitted · 0 rank-eligible
Function and tool calling
PreviewEmit correct schema-valid tool calls
1 implemented · 0 admitted · 0 rank-eligible
MCP tool orchestration
UnavailableCoordinate MCP tools to complete a task
0 implemented · 0 admitted · 0 rank-eligible
Web browsing and navigation
UnavailableRetrieve and act on live web content
0 implemented · 0 admitted · 0 rank-eligible
Computer use
UnavailableOperate a graphical interface to complete a task
0 implemented · 0 admitted · 0 rank-eligible
Deep research
UnavailableSynthesize multiple sources with traceable citations
0 implemented · 0 admitted · 0 rank-eligible
Customer support agent
UnavailableResolve user support issues end to end
0 implemented · 0 admitted · 0 rank-eligible
Enterprise and CRM workflow
UnavailableExecute business workflows across enterprise systems
0 implemented · 0 admitted · 0 rank-eligible
Mathematical reasoning
UnavailableSolve quantitative or symbolic problems
0 implemented · 0 admitted · 0 rank-eligible
General knowledge QA
UnavailableAnswer broad knowledge questions accurately
0 implemented · 0 admitted · 0 rank-eligible
RAG and retrieval
UnavailableRetrieve relevant context and ground an answer
0 implemented · 0 admitted · 0 rank-eligible
Long-term memory
UnavailablePersist and recall useful information across sessions
0 implemented · 0 admitted · 0 rank-eligible
Finance
UnavailablePerform domain-grounded financial reasoning and workflows
0 implemented · 0 admitted · 0 rank-eligible
Legal
UnavailablePerform domain-grounded legal reasoning and drafting
0 implemented · 0 admitted · 0 rank-eligible
Medical
UnavailablePerform domain-grounded clinical reasoning and question answering
0 implemented · 0 admitted · 0 rank-eligible
Multilingual
UnavailableMaintain quality across languages and translation tasks
0 implemented · 0 admitted · 0 rank-eligible
Vision and multimodal
UnavailableReason over images, audio, or video
0 implemented · 0 admitted · 0 rank-eligible
Web frontend code generation
UnavailableBuild and iterate web interfaces from a specification
0 implemented · 0 admitted · 0 rank-eligible
SRE incident response
PreviewDiagnose and repair live service or infrastructure incidents
1 implemented · 0 admitted · 0 rank-eligible
DevOps lifecycle
UnavailableBuild, configure, test, deploy, and monitor software delivery systems
0 implemented · 0 admitted · 0 rank-eligible
Terminal generalist
PreviewComplete heterogeneous computer tasks through a terminal
1 implemented · 0 admitted · 0 rank-eligible
Mobile app code generation
UnavailableBuild and iterate native or cross-platform mobile apps
0 implemented · 0 admitted · 0 rank-eligible
Reasoning
PreviewSolve novel multi-step reasoning tasks
1 implemented · 0 admitted · 0 rank-eligible
Factuality
UnavailableProduce correct and grounded factual claims
0 implemented · 0 admitted · 0 rank-eligible
Professional deliverables
PreviewCreate review-ready professional work products from a complete brief, domain context, and reference files.
1 implemented · 0 admitted · 0 rank-eligible
Machine-learning engineering
UnavailableBuild, train, and optimize machine-learning solutions from datasets and scored task objectives.
0 implemented · 0 admitted · 0 rank-eligible
Computational research reproduction
UnavailableReproduce published computational results by implementing or executing experiments from papers, code, data, and environments.
0 implemented · 0 admitted · 0 rank-eligible
Unavailable is not a verdict that a benchmark is dead. It means no implemented feed currently supports a public ranking; source recovery and scraper work can still change that status.