OpenGPTEvidence-based model decisionsAgent categories
Research AI agents
This category contains 9 agent products whose identity is supported by a first-party product page, documentation page, or repository.
- Verified products
- 9
- Published benchmarks
- 1
- Published rows
- 15
- Identity reviewed
Evidence scope
This page is an index of existing public evidence and provider-documented identities. Inclusion does not imply an OpenGPT endorsement or a universal rank. Research.
How OpenGPT handles this evidence
Agent product identity and benchmark evidence stay separate. Product inclusion requires a first-party source; ranking rows retain the full model, scaffold, tools, budget, and environment published by the benchmark.
Comparison boundaries
- Product inclusion is not a benchmark rank or endorsement.
- Only configurations from the same benchmark version and scope should be compared.
- A benchmarked configuration does not establish the quality of every version of the product.
Verified products
This category contains 9 agent products whose identity is supported by a first-party product page, documentation page, or repository.
- Claude ResearchAnthropicresearch agentcommercialIdentity reviewed: 2026-08-02
- Gemini Deep Research AgentGoogleresearch agentcommercialIdentity reviewed: 2026-08-02
- Glean Deep ResearchGleanresearch agentcommercialIdentity reviewed: 2026-08-02
- GPT ResearcherGPT Researcheropen source agentopen sourceIdentity reviewed: 2026-08-02
- HAL Generalist AgentHolistic Agent Leaderboardresearch agentresearch previewIdentity reviewed: 2026-08-02
- Open Deep ResearchHugging Faceresearch agentopen sourceIdentity reviewed: 2026-08-02
- OpenAI deep researchOpenAIresearch agentcommercialIdentity reviewed: 2026-08-02
- Perplexity ResearchPerplexityresearch agentcommercialIdentity reviewed: 2026-08-02
- RobinFutureHouseresearch agentopen sourceIdentity reviewed: 2026-08-02