Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 16 additions & 0 deletions cli/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,22 @@

---

## New: PipelineScore v4

v4 is the current board: 71 tasks across code, repo bug fixes, multi-step agent work, tool calls, reasoning, long documents and instructions. Model-written code runs in a Docker sandbox (no network), and every answer is checked by a program. Docker must be running.

```bash
npx @pipelinescore/cli v4 \
--provider local \
--endpoint http://localhost:11434/v1 \
--model qwen3.8:27b \
--save my-run.json

npx @pipelinescore/cli v4-submit my-run.json # post it to pipelinescore.ai/leaderboard
```

Every provider gets the same answer budget, and `v4-submit` refuses runs that stopped early or hit provider errors. The 34-task v3 commands below still work; their board lives at [pipelinescore.ai/leaderboard/v3](https://pipelinescore.ai/leaderboard/v3).

## Quickstart — zero config

```bash
Expand Down
4 changes: 2 additions & 2 deletions cli/package-lock.json

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

4 changes: 2 additions & 2 deletions cli/package.json
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
{
"name": "@pipelinescore/cli",
"version": "0.4.2",
"description": "PipelineScore CLI \u2014 benchmark LLMs on your own hardware. Zero-config: `npx @pipelinescore/cli` auto-detects Ollama / LM Studio / llama.cpp / MLX and walks you through a run. Fully deterministic suite (no API key required), scored locally, published to the hardware-aware public leaderboard at pipelinescore.ai.",
"version": "0.5.0",
"description": "PipelineScore CLI — benchmark LLMs on your own hardware. Zero-config: `npx @pipelinescore/cli` auto-detects Ollama / LM Studio / llama.cpp / MLX and walks you through a run. Fully deterministic suite (no API key required), scored locally, published to the hardware-aware public leaderboard at pipelinescore.ai.",
"type": "module",
"main": "./dist/index.js",
"bin": {
Expand Down
Loading