Catalog Systems Use-case-specific agent harnesses OpenAI Deep Research

OpenAI Deep Research

https://openai.com/index/introducing-deep-research/

Cross-listing: OpenAI Deep Research, launched Feb-2025 inside ChatGPT, is OpenAI's flagship long-horizon research agent (o3 base model). Distinct from the OpenAI Codex / Computer Use rows because the focus is web-research + multi-step synthesis. Available to Pro / Plus / Team / Enterprise users. The separate API models o3-deep-research and o4-mini-deep-research were retired on 2026-07-23, with gpt-5.6-sol as OpenAI's recommended replacement.

Active · Verified 2026-10-31 · fresh as of 2026-10-31 · 43 cited facts, 0 with a third-party source · Suggest a correction

At a glance

Type
Autonomous research agent (cross-listing harness) vendor fix
Tier
T1
Section
Use-case-specific agent harnesses
Created
2025-02 (Deep Research launched) vendor fix
Latest release
not applicable — wrong section fix
License
not applicable — not OSS fix
GitHub
not applicable — no GitHub repo fix
Pricing
Pro $200/mo (100 queries/mo); Plus tier 10 queries/mo (post-Apr 2025) vendor fix
Funding
not applicable — wrong section fix
Freshness
2026-10-31 fix
Layer
agent-infra repo fix
Licensing
commercial vendor fix
Deployment locus
cloud vendor fix
Platform vs product
product vendor fix
SDK languages
csharp, go, java, python, rest, ruby, typescript vendor fix
Protocols
mcp-client vendor fix
Compliance
gdpr, soc2-type2 vendor fix

When to use

Optimised for: autonomous web research + synthesis

Anti-fit: not applicable — wrong section

Pros & cons

Pros

Highest-quality research-agent in production as of 2025 H1; HLE benchmark dominance; integrated with ChatGPT distribution.

Cons

ChatGPT-only; Pro-tier-first ($200/mo); rate-limited; long-horizon turn-time (10–30 min).

Claims & capabilities

Launched Feb-2025 in ChatGPT (Pro tier first); built on o3; tackles 30-min+ multi-step research; benchmarks: HLE (Humanity's Last Exam) 26.6% (vs GPT-4o 3.3%)

Technical surface

API surface
not applicable — wrong section
Backend storage
not applicable — wrong section
Deployment
not applicable — wrong section
Embedding model
not applicable — not a memory product
Multi-tenancy
not applicable — wrong section
MCP
not applicable — wrong section
A2A
not applicable — wrong section
OpenTelemetry
not applicable — wrong section

Alternatives to OpenAI Deep Research → live options ranked by shared design, licence, hosting and MCP support

Similar systems

Other use-case-specific agent harnesses in the catalog, ranked by inbound references.

  • Calypso AI (now F5 AI Guardrails) T2

    Dublin / NYC founded 2018; GenAI security and content scanning. $23M Series A-1 Aug-2024 (Paladin Capital + Lockheed Martin Ventures). Pivoted from defence ML to enterprise GenAI guardrails 2023-24.

  • Salesforce Agentforce T1

    Cross-listing: Salesforce Agentforce already exists under Vertical / Customer-support / voice memory; this row is the Agent-harness framing — Agentforce as the consumer-facing agent IDE / builder on Data Cloud. Agentforce launched Sep-2024; Agentforce 2.0 announced Dec-2024.

  • Lasso Security T2

    Tel-Aviv-founded LLM security platform — adversarial testing, prompt-injection defence, data-leakage prevention. $6M seed Nov-2023 (Entrée Capital + Samsung Next).

  • Pillar Security T2

    Tel-Aviv-founded (Dor Sarig + Ziv Karliner, both ex-IDF cyber); end-to-end LLM security platform — discovery, red-teaming, runtime guardrails. $9M seed Apr-2024 (Shield Capital + Greenfield).

  • Promptfoo T2

    OSS (MIT) prompt-engineering eval and red-team CLI; supports OWASP LLM Top-10 vulnerability tests, jailbreak generation, and CI/CD pipeline integration. $5M seed Mar-2024 (Andreessen Horowitz). Founders ex-Discord + ex-Smile Identity.

  • Protect AI T2

    Founded 2022 by Ian Swanson + Daryan Dehghanpisheh (ex-AWS); AI/ML supply-chain + GenAI security. Acquired by Palo Alto Networks (2025-07-22; announced Apr-2025, reported $700M); its capabilities now ship in Prisma AIRS. Products: Recon (red-team), ModelScan (model artifact scanner), Sightline (vulnerability DB).

Sources (42 cited facts, 0 third-party)

Each fact links to the page it came from; the badge says who published it (how to read sources).