Promptfoo 0.124.0
Open-source LLM testing, red-teaming, and vulnerability-scanning CLI, updated in v0.124.0 with Claude Opus 5.5 support and new filtering options.
Overview
Promptfoo, the open-source CLI for testing prompts, agents, and RAG pipelines plus red-teaming/vulnerability scanning for AI apps, released v0.124.0 on October 6, 2026. The update adds Claude Opus 5.5 model support and comma-separated OR values in --filter-metadata, while removing the hosted ChatKit provider and making the WatsonX, Langfuse, and Slack SDKs opt-in dependencies to slim the default install, alongside fixes to CSV handling and evaluation workflows.
This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.
Key features
- Declarative config-based testing of prompts, agents, and RAG pipelines, with CLI and CI/CD integration
- Red-teaming and vulnerability scanning against OWASP-style LLM risk categories
- Compares outputs across GPT, Claude, Gemini, DeepSeek, and other model providers
- v0.124.0: adds Claude Opus 5.5 support, comma-separated OR filters in --filter-metadata
- v0.124.0: ChatKit hosted provider removed; WatsonX, Langfuse, and Slack SDKs made opt-in
Use cases
- CI/CD gating on prompt/agent regression tests before deployment
- Red-team scanning of LLM applications for security vulnerabilities
- Side-by-side evaluation of multiple model providers on the same test suite
Pricing
Open source (free) CLI; Promptfoo also offers a hosted/enterprise tier, pricing not disclosed in this release's coverage.
Pricing reflects the source record and may change. Confirm current terms with the provider.
Dates and sources
- Released
- 6 Oct 2026
- Record updated
- 12 Oct 2026
- Record last checked
- 12 Oct 2026