Add Terminal Bench 2.0 evaluation result

.eval_results/tmax-2b.yaml

@@ -0,0 +1,19 @@
+# Evaluation results for allenai/tmax-2b
+# Extracted from the model card evaluation table
+# https://huggingface.co/allenai/tmax-2b
+# Paper: https://arxiv.org/abs/2606.23321
+
+# ---------------------------------------------------------------------------
+# Terminal Bench (TB 2.0 / daytona) - registered on Hub
+# ---------------------------------------------------------------------------
+
+- dataset:
+    id: harborframework/terminal-bench-2.0
+    task_id: terminalbench_2
+  value: 2.9
+  date: "2026-06-23"
+  source:
+    url: https://huggingface.co/allenai/tmax-2b
+    name: Tmax 2B model card evaluation table
+  notes: "TB 2.0 (daytona) score. Mean ± std error over 5 runs."
+