Files
goose/.github/workflows/terminal-benchmarks.yml
Michael Neale 0ffb576597 version
2025-11-19 16:26:12 +11:00

34 lines
841 B
YAML

on:
push:
branches:
- benchmarks
workflow_dispatch:
name: TBench
jobs:
run-tbench:
name: Run TBench
runs-on: ubuntu-latest
steps:
- name: Checkout Code
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # pin@v4
- name: Set up Python
uses: actions/setup-python@0b93645e9fea7318ecaed2b359559ac225c90a2b # pin@v5
with:
python-version: '3.13'
- name: Install uv
run: |
curl -LsSf https://astral.sh/uv/install.sh | sh
echo "$HOME/.cargo/bin" >> $GITHUB_PATH
- name: Install Harbor
run: uv tool install harbor
- name: Run TBench
run: harbor run -d terminal-bench@2.0 --agent goose --model claude-sonnet-4-5-20250929
env:
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}