Add community evaluation results for SWE-BENCH_VERIFIED

This PR adds community-provided evaluation results for the following benchmarks:

- **[SWE-BENCH_VERIFIED](https://huggingface.co/datasets/SWE-bench/SWE-bench_Verified)**

These results were extracted from the model card. This is based on the new [evaluation results feature](https://huggingface.co/docs/hub/eval-results).

*Note: This is an automated PR. Please review the evaluation results before merging.*

.eval_results/swe-bench_verified.yaml

@@ -0,0 +1,7 @@
+- dataset:
+    id: SWE-bench/SWE-bench_Verified
+    task_id: swe_bench_%_resolved
+  value: 77.6
+  source:
+    url: https://huggingface.co/mistralai/Mistral-Medium-3.5-128B
+    name: Model Card