PawBench 🐾

    / \__
   (    @\___    PawBench
  /         O   4-dimensional LLM inference benchmark
 /   (_____/    "More bark than bite"
/_____/   U

Because your model deserves a benchmark with more bark than bite.

4-dimensional LLM inference benchmark for OpenAI-compatible endpoints. Multi-turn, multi-agent, parallel dispatch with tool calling.

Tests your model with realistic coding agent workloads — not synthetic single-turn completions.

Meet Lola

PawBench is inspired by Lola (@_justlolathings) — the most fashionable pup on Instagram. The built-in scenarios revolve around building her boutique dog apparel store, PawStyle by Lola. Every product, every size guide, every "Lola's Pick" badge traces back to this style icon on four legs.

Follow Lola: https://www.instagram.com/_justlolathings/

Install

pip install pawbench
# or
uv pip install pawbench

Quick Start

# Benchmark your local vLLM
pawbench --endpoint http://localhost:8000

# Against any OpenAI-compatible endpoint
pawbench --endpoint https://api.openai.com/v1 --tag gpt4o

# Just throughput saturation (no scenarios)
pawbench --saturation-only --concurrency 1,2,4,8,16

# JSON output for CI/autoresearch
pawbench --json --output results/

# Custom scenario
pawbench --scenario my_scenario.json

What It Measures

4 Dimensions

Built-in Scenarios: PawStyle by Lola

Two parallel agents build Lola's boutique dog apparel e-commerce store — "Where every pup is a fashionista":

• pawstyle-independent — Frontend and backend work independently on Lola's shop. Pure parallel throughput + quality baseline.
• pawstyle — Backend gets a steering event mid-task ("frontend added a Size Guide button — implement Lola's breed-specific sizing endpoint").
• pawstyle-nudge — Frontend adds Lola's Favorites (wishlist) and Compare features that require backend changes. Backend receives nudges and adapts.

Each scenario is 3 turns x 2 agents, with tool calls (write_file, read_file, run_command) and injected tool results. Products include Lola's Signature Bandana, Cozy Knit Sweater, Rainy Day Raincoat, Adventure Booties, Dapper Bow Tie, and Walk-in-Style Harness — with "Lola's Pick" badges on her personal favorites.

Server Metrics (optional)

If the endpoint exposes /metrics (vLLM, TGI), PawBench scrapes:

• KV cache usage and prefix cache hit rate
• Speculative decoding acceptance rate
• GPU cache pressure

Custom Scenarios

Scenarios are JSON files:

{
  "id": "my-scenario",
  "name": "My Custom Scenario",
  "agents": [
    {
      "id": "agent-1",
      "name": "My Agent",
      "turns": [
        {
          "turn": 1,
          "role": "user",
          "content": "Build a REST API with Flask...",
          "tools": ["write_file"],
          "expect": {
            "tool_calls_min": 1,
            "tool_name_any": ["write_file"],
            "output_mentions": ["flask", "api"]
          }
        }
      ]
    }
  ],
  "tools_schema": [...]
}

Comparing Configs

pawbench --tag baseline --output results/
# ... change model config ...
pawbench --tag eagle3 --output results/

python -m pawbench.compare results/pawbench_baseline_*.json results/pawbench_eagle3_*.json

Output Format

JSON results include full model card (architecture, quantization, GPU, serving params) for reproducibility:

{
  "tag": "fp8-eagle3-spec3",
  "model_card": {
    "model_name": "qwen3-coder",
    "model_config": {"architectures": ["Qwen3NextForCausalLM"], "num_experts": 512, "...": "..."},
    "tuning": {"kv_cache_dtype": "fp8_e4m3", "speculative_config": "eagle3", "...": "..."},
    "gpu": {"name": "NVIDIA GB10", "...": "..."}
  },
  "dim1_throughput": {"avg_single_tok_s": 69.0, "raw_peak_tok_s": 469.3, "...": "..."},
  "dim2_quality": {"avg_quality": 0.81, "tool_accuracy": 0.96, "...": "..."},
  "saturation_curve": [{"concurrency": 1, "tok_s": 69.3}, {"concurrency": 8, "tok_s": 469.3}],
  "server_metrics": {"spec_acceptance_rate": 0.72, "gpu_prefix_cache_hit_rate": 0.92}
}

License

MIT