diff --git a/.github/workflows/kani.yml b/.github/workflows/kani.yml
index e40484f852ce..2f19a5fb2c7d 100644
--- a/.github/workflows/kani.yml
+++ b/.github/workflows/kani.yml
@@ -306,4 +306,20 @@ jobs:
       - name: Run benchcomp
         run: |
           new/tools/benchcomp/bin/benchcomp \
-            --config new/tools/benchcomp/configs/perf-regression.yaml
+            --config new/tools/benchcomp/configs/perf-regression.yaml \
+            run
+          new/tools/benchcomp/bin/benchcomp \
+            --config new/tools/benchcomp/configs/perf-regression.yaml \
+            collate
+
+      - name: Perf Regression Results Table
+        run: |
+          new/tools/benchcomp/bin/benchcomp \
+            --config new/tools/benchcomp/configs/perf-regression.yaml \
+            visualize --only dump_markdown_results_table >> "$GITHUB_STEP_SUMMARY"
+
+      - name: Run other visualizations
+        run: |
+          new/tools/benchcomp/bin/benchcomp \
+            --config new/tools/benchcomp/configs/perf-regression.yaml \
+            visualize --except dump_markdown_results_table
diff --git a/tools/benchcomp/benchcomp/visualizers/__init__.py b/tools/benchcomp/benchcomp/visualizers/__init__.py
index 7fa2b3b750ee..c4870c425f19 100644
--- a/tools/benchcomp/benchcomp/visualizers/__init__.py
+++ b/tools/benchcomp/benchcomp/visualizers/__init__.py
@@ -3,7 +3,9 @@
 
 
 import dataclasses
+import textwrap
 
+import jinja2
 import yaml
 
 import benchcomp
@@ -76,3 +78,73 @@ def __call__(self, results):
         with self.get_out_file() as handle:
             print(
                 yaml.dump(results, default_flow_style=False), file=handle)
+
+
+
+class dump_markdown_results_table:
+    """Print a Markdown-formatted table displaying benchmark results
+
+    The 'out_file' key is mandatory; specify '-' to print to stdout.
+
+    Sample configuration:
+
+    visualize:
+    - type: dump_markdown_results_table
+      out_file: '-'
+    """
+
+
+    def __init__(self, out_file):
+        self.get_out_file = benchcomp.Outfile(out_file)
+
+
+    @staticmethod
+    def _get_template():
+        return textwrap.dedent("""\
+            {% for metric, benchmarks in d["metrics"].items() %}
+            ## {{ metric }}
+
+            | Benchmark | {% for variant in d["variants"] %} {{ variant }} |{% endfor %}
+            | --- | {% for variant in d["variants"] %}--- |{% endfor -%}
+            {% for bench_name, bench_variants in benchmarks.items () %}
+            | {{ bench_name }} {% for variant in d["variants"] -%}
+             | {{ bench_variants[variant] }} {% endfor %}|
+            {%- endfor %}
+            {% endfor -%}
+            """)
+
+
+    @staticmethod
+    def _get_variant_names(results):
+        return results.values()[0]["variants"]
+
+
+    @staticmethod
+    def _organize_results_into_metrics(results):
+        ret = {metric: {} for metric in results["metrics"]}
+        for bench, bench_result in results["benchmarks"].items():
+            for variant, variant_result in bench_result["variants"].items():
+                for metric, value in variant_result["metrics"].items():
+                    try:
+                        ret[metric][bench][variant] = variant_result["metrics"][metric]
+                    except KeyError:
+                        ret[metric][bench] = {
+                            variant: variant_result["metrics"][metric]
+                    }
+        return ret
+
+
+    def __call__(self, results):
+        data = {
+            "metrics": self._organize_results_into_metrics(results),
+            "variants": list(results["benchmarks"].values())[0]["variants"],
+        }
+
+        env = jinja2.Environment(
+            loader=jinja2.BaseLoader, autoescape=jinja2.select_autoescape(
+                enabled_extensions=("html"),
+                default_for_string=True))
+        template = env.from_string(self._get_template())
+        output = template.render(d=data)[:-1]
+        with self.get_out_file() as handle:
+            print(output, file=handle)
diff --git a/tools/benchcomp/configs/perf-regression.yaml b/tools/benchcomp/configs/perf-regression.yaml
index cb9d0c493fb4..a76cb5ab54bf 100644
--- a/tools/benchcomp/configs/perf-regression.yaml
+++ b/tools/benchcomp/configs/perf-regression.yaml
@@ -29,7 +29,10 @@ run:
 
 visualize:
   - type: dump_yaml
-    out_file: '/tmp/result.yaml'
+    out_file: '-'
+
+  - type: dump_markdown_results_table
+    out_file: '-'
 
   - type: error_on_regression
     variant_pairs: [[kani_old, kani_new]]
diff --git a/tools/benchcomp/test/test_regression.py b/tools/benchcomp/test/test_regression.py
index 3e6afb832662..258005f84b57 100644
--- a/tools/benchcomp/test/test_regression.py
+++ b/tools/benchcomp/test/test_regression.py
@@ -8,6 +8,7 @@
 import pathlib
 import subprocess
 import tempfile
+import textwrap
 import unittest
 
 import yaml
@@ -391,6 +392,69 @@ def test_error_on_regression_visualization_ratio_regressed(self):
                 run_bc.proc.returncode, 1, msg=run_bc.stderr)
 
 
+    def test_markdown_results_table(self):
+        """Run the markdown results table visualization"""
+
+        with tempfile.TemporaryDirectory() as tmp:
+            run_bc = Benchcomp({
+                "variants": {
+                    "variant_1": {
+                        "config": {
+                            "directory": str(tmp),
+                            "command_line":
+                                "mkdir bench_1 bench_2"
+                                "&& echo true > bench_1/success"
+                                "&& echo true > bench_2/success"
+                                "&& echo 5 > bench_1/runtime"
+                                "&& echo 10 > bench_2/runtime"
+                        },
+                    },
+                    "variant_2": {
+                        "config": {
+                            "directory": str(tmp),
+                            "command_line":
+                                "mkdir bench_1 bench_2"
+                                "&& echo true > bench_1/success"
+                                "&& echo false > bench_2/success"
+                                "&& echo 10 > bench_1/runtime"
+                                "&& echo 5 > bench_2/runtime"
+                        }
+                    }
+                },
+                "run": {
+                    "suites": {
+                        "suite_1": {
+                            "parser": { "module": "test_file_to_metric" },
+                            "variants": ["variant_1", "variant_2"]
+                        }
+                    }
+                },
+                "visualize": [{
+                    "type": "dump_markdown_results_table",
+                    "out_file": "-",
+                }]
+            })
+            run_bc()
+
+            self.assertEqual(run_bc.proc.returncode, 0, msg=run_bc.stderr)
+            self.assertEqual(
+                run_bc.stdout, textwrap.dedent("""
+                    ## runtime
+
+                    | Benchmark |  variant_1 | variant_2 |
+                    | --- | --- |--- |
+                    | bench_1 | 5 | 10 |
+                    | bench_2 | 10 | 5 |
+
+                    ## success
+
+                    | Benchmark |  variant_1 | variant_2 |
+                    | --- | --- |--- |
+                    | bench_1 | True | True |
+                    | bench_2 | True | False |
+                    """))
+
+
     def test_only_dump_yaml(self):
         """Ensure that benchcomp terminates with return code 0 when `--only dump_yaml` is passed, even if the error_on_regression visualization would have resulted in a return code of 1"""