feat(benchmarks): add model matrix run script

benchmarks/run.sh

@@ -0,0 +1,44 @@
+#!/usr/bin/env bash
+
+set -euo pipefail
+
+scenario_dir="${1:-${SCENARIO_DIR:-}}"
+leaderboard_dir="${2:-${LEADERBOARD_DIR:-}}"
+
+if [[ -z "$scenario_dir" || -z "$leaderboard_dir" ]]; then
+  printf 'Usage: %s SCENARIO_DIR LEADERBOARD_DIR\n' "$0" >&2
+  printf '       or set SCENARIO_DIR and LEADERBOARD_DIR\n' >&2
+  exit 2
+fi
+
+agent_name=stirrup_agent
+scenario_ids=lite
+
+model_configs=(
+  "litellm_proxy/gcp/gemini-3.6-flash high"
+  "litellm_proxy/azure/gpt-5.6-sol max"
+  "tokenrouter/z-ai/glm-5.2 max"
+  "tokenrouter/MiniMax-M3 high"
+  "litellm_proxy/aws/claude-opus-5 high"
+  "litellm_proxy/azure/DeepSeek-V4-Flash max"
+  "litellm_proxy/aws/gpt-oss-120b high"
+  "litellm_proxy/aws/claude-sonnet-5 max"
+)
+
+for model_config in "${model_configs[@]}"; do
+  read -r model_id reasoning_effort <<< "$model_config"
+
+  echo "Running $model_id with reasoning effort $reasoning_effort"
+
+  uv run python -m benchmark.scenario_suite_runner \
+    --scenario-ids "$scenario_ids" \
+    --scenario-root "$scenario_dir" \
+    --agent_name "$agent_name" \
+    --model-id "$model_id" \
+    --reasoning-effort "$reasoning_effort" \
+    --trajectory-root "$leaderboard_dir/assetopsbench-trajectories" \
+    --reports-root "$leaderboard_dir/assetopsbench-reports" \
+    --stirrup-workspace-root "$leaderboard_dir/assetopsbench-stirrup-workspaces" \
+    --skip-existing \
+    --continue-on-error
+done