windmill/ai_evals

AI Evals

Small benchmark runner for the four Windmill AI generation modes:

• cli
• flow
• script
• app

The benchmark always tests the current production prompts, tools, and guidance in this checkout.

Each attempt runs:

1. the real production path
2. deterministic validation
3. LLM judging

Install

cd ai_evals
bun install

Frontend modes also require frontend dependencies:

cd frontend
bun install

Commands

List model aliases:

cd ai_evals
bun run cli -- models

List cases:

cd ai_evals
bun run cli -- cases
bun run cli -- cases flow

Run benchmarks:

cd ai_evals
bun run cli -- run flow
bun run cli -- run flow flow-test4-order-processing-loop --model opus
bun run cli -- run flow flow-test0-sum-two-numbers --models haiku,opus,4o
bun run cli -- run flow flow-test0-sum-two-numbers --runs 3 --verbose
bun run cli -- run flow --record
WMILL_AI_EVAL_BACKEND_URL=http://127.0.0.1:8000 bun run cli -- run flow --backend-validation preview
bun run cli -- run cli bun-hello-script

Public CLI surface:

• models
• cases [mode]
• run <mode> [caseIds...]

run options:

• --runs <n>: repeat each case n times
• --output <path>: custom result JSON path
• --model <alias>: choose the model under test
• --models <a,b,c>: run the same cases sequentially against several model aliases
• --verbose: stream assistant output for frontend runs
• --record: append a compact tracked summary line to ai_evals/history/<mode>.jsonl for full-suite runs only
• --backend-validation <mode>: optional backend smoke validation (off or preview) for script and flow evals

Models

Use bun run cli -- models to see the current aliases.

Today:

• haiku
• sonnet
• opus
• 4o
• gemini-flash
• gemini-pro
• gemini-3-flash-preview
• gemini-3.1-pro-preview

Notes:

• the command also prints accepted alias spellings such as gpt-4o, claude-opus-4.6, and claude-haiku-4.5
• frontend modes (flow, script, app) can use Anthropic, OpenAI, and Gemini-backed aliases
• cli mode always uses the Anthropic agent SDK, so only Anthropic aliases are valid there
• the judge model is separate and currently defaults to claude-sonnet-4-6

Case Format

Cases live in one YAML file per mode under ai_evals/cases/.

Minimal shape:

- id: flow-test0-sum-two-numbers
  prompt: |-
    Create a flow that takes two numbers, `a` and `b`, and returns their sum.
  initial: ai_evals/fixtures/...
  expected: ai_evals/fixtures/...

Optional fields:

• initial: starting state fixture
• expected: expected artifact fixture
• validate: extra deterministic validation rules
• runtime.backendPreview: optional real backend preview config for smoke validation

For flow mode, validate can express requirements such as:

• accepted input schema shapes
• required results.* reference validity
• required module/code/input characteristics

For flow mode, an initial fixture can also include a benchmark workspace catalog of existing scripts and flows. That lets the real search_workspace and get_runnable_details tools discover reusable workspace runnables during evals.

If --backend-validation preview is enabled:

• script evals run a real backend script preview in an isolated temp workspace
• flow evals run a real backend flow preview only for cases that define runtime.backendPreview
• flow cases with initial.workspace fixtures seed those scripts and flows into the preview workspace before preview
• when WMILL_AI_EVAL_BACKEND_WORKSPACE is set, ai_evals treats that workspace as a dedicated test workspace, clears managed eval assets under f/evals/* before each preview run, and then reseeds the current case fixtures

Supported backend validation env vars:

• WMILL_AI_EVAL_BACKEND_VALIDATION=preview
• WMILL_AI_EVAL_BACKEND_URL=http://127.0.0.1:8000
• WMILL_AI_EVAL_BACKEND_EMAIL=admin@windmill.dev
• WMILL_AI_EVAL_BACKEND_PASSWORD=changeme
• WMILL_AI_EVAL_BACKEND_WORKSPACE=integration-tests to reuse an existing workspace on CE installs with low workspace limits
• WMILL_AI_EVAL_KEEP_WORKSPACES=1
• WMILL_AI_EVAL_WORKSPACE_PREFIX=ai-evals

Results And Artifacts

Every run writes:

• a summary JSON under ai_evals/results/
• generated artifacts in a sibling directory

If --record is used, the CLI also appends one compact JSON line to:

• ai_evals/history/flow.jsonl
• ai_evals/history/script.jsonl
• ai_evals/history/app.jsonl
• ai_evals/history/cli.jsonl

Each recorded line contains:

• run metadata (createdAt, gitSha, mode, runModel, judgeModel)
• suite totals (caseCount, attemptCount, passedAttempts, passRate, averageDurationMs, averageJudgeScore)
• average token usage (averageTokenUsagePerAttempt)
• per-case metrics under cases[] (averageDurationMs, averageJudgeScore, averageTokenUsagePerAttempt, pass rate)
• failedCaseIds

Example:

• summary: ai_evals/results/2026-04-09T09-40-33.051Z__flow.json
• artifacts: ai_evals/results/2026-04-09T09-40-33.051Z__flow/

Typical artifacts by mode:

• flow: flow.json
• script: script.json plus the generated script file
• app: app.json plus frontend/backend files
• cli: assistant-output.txt plus generated workspace files
• backend-validated attempts also include backend-preview.json

Layout

• cases/: one YAML file per mode
• fixtures/: initial and expected fixtures
• core/: shared loading, model resolution, validation, judging, and result writing
• modes/: one runner per mode
• history/: optional tracked pass-rate history written by run --record, one JSONL file per mode
• results/: local benchmark output and artifacts

Notes

• Frontend modes reuse the production frontend chat code through the Vitest bridge.
• CLI mode creates an isolated workspace, writes the current checkout guidance into it, and benchmarks the real skills / AGENTS.md flow.
• Frontend progress streams live while the benchmark is running.
• Deterministic validators should stay focused on real correctness constraints, not one exact implementation shape.