diff --git a/dataset_builder/humaneval_to_el.py b/dataset_builder/humaneval_to_el.py
new file mode 100644
index 0000000000..8b70d095d6
--- /dev/null
+++ b/dataset_builder/humaneval_to_el.py
@@ -0,0 +1,66 @@
+"""
+This script translates problems from the OpenAI HumanEval dataset into Emacs Lisp.
+
+- Home: https://www.gnu.org/software/emacs/
+- Reference manual: https://www.gnu.org/software/emacs/manual/elisp.html
+- Test library: https://www.gnu.org/software/emacs/manual/ert.html
+"""
+
+import ast
+from typing import List
+
+
+class Translator:
+
+    USub = "-"
+
+    stop = ["\n(defun", "\n;", "\n("]
+
+    def file_ext(self):
+        return "el"
+
+    def translate_prompt(
+        self, name: str, args: List[ast.arg], _returns, description: str
+    ) -> str:
+        el_preamble = ";;; -*- lexical-binding: t; -*-"
+        el_args = " ".join([arg.arg for arg in args])
+        el_description = description.replace('"', '\\"')
+        self.entry_point = name
+        return f'{el_preamble}\n(defun {name} ({el_args})\n"{el_description}"\n'
+
+    def test_suite_prefix_lines(self, entry_point) -> List[str]:
+        return [
+            f"(defalias #'candidate #'{entry_point})",
+            "(ert-deftest test-human-eval ()",
+        ]
+
+    def test_suite_suffix_lines(self) -> List[str]:
+        return [")"]
+
+    def deep_equality(self, left: str, right: str) -> str:
+        return f"  (should (equal {left} {right}))"
+
+    def gen_literal(self, c: bool | str | int | float):
+        if type(c) is bool:
+            return "t" if c else "nil"
+        elif type(c) is str:
+            return f'"{c}"'
+        elif c is None:
+            return "nil"
+        return repr(c)
+
+    def gen_var(self, variable: str) -> str:
+        return variable
+
+    def gen_list(self, list: List[str]) -> str:
+        return "(list " + " ".join(list) + ")"
+
+    def gen_tuple(self, tuple: List[str]) -> str:
+        return "(list " + " ".join(tuple) + ")"
+
+    def gen_dict(self, keys: List[str], values: List[str]) -> str:
+        pairs = " ".join(f"(cons {k} {v})" for k, v in zip(keys, values))
+        return "(list " + pairs + ")"
+
+    def gen_call(self, func: str, args: List[str]) -> str:
+        return "(" + func + " " + " ".join(args) + ")"
diff --git a/dataset_builder/libexperiments.py b/dataset_builder/libexperiments.py
index c896e6024b..5e2654b39f 100644
--- a/dataset_builder/libexperiments.py
+++ b/dataset_builder/libexperiments.py
@@ -42,14 +42,15 @@ def path(self) -> Path:
     "elixir",
     "clj",
     "ada",
+    "el",
 ]
 MODELS = ["davinci", "incoder", "codegen"]
 
 def all_experiments() -> Iterator[Experiment]:
     """
     An iterator that produces (lang, model, temp, variation) tuples for all
-    the standard experiments that we care about. The ../experiments directory 
-    has results from configurations that were explored and determined 
+    the standard experiments that we care about. The ../experiments directory
+    has results from configurations that were explored and determined
     uninteresting for a full result. (We are not deleting results.)
     """
     for dataset in DATASETS:
diff --git a/dataset_builder/terms.csv b/dataset_builder/terms.csv
index 8b089d40fe..7eb3950b8d 100644
--- a/dataset_builder/terms.csv
+++ b/dataset_builder/terms.csv
@@ -26,3 +26,4 @@ Matlab,m,array,array,array,dictionary,<missing>,true,false
 Haskell,hs,list,list,tuple,association list,Nothing,True,False
 Clojure,clj,vector,list,vector,map,nil,true,false
 Dart,dart,list,list,record,map,null,true,false
+Emacs Lisp,el,list,list,list,alist,nil,t,nil
diff --git a/evaluation/Dockerfile b/evaluation/Dockerfile
index 0808bfb027..e8b85362bf 100644
--- a/evaluation/Dockerfile
+++ b/evaluation/Dockerfile
@@ -106,6 +106,9 @@ RUN apt-get update -yqq && apt-get install -yqq dart
 # Lean
 # RUN wget https://github.com/leanprover/lean4/releases/download/v4.6.0-rc1/lean-4.6.0-rc1-linux.zip -O /tmp/lean.zip && unzip /tmp/lean.zip -d /root/lean/ && ln -s /root/lean/bin/lean /bin/lean
 
+# Emacs Lisp (no-X/GUI version)
+RUN apt install -y emacs-nox
+
 # install numpy for humanevalplus
 RUN python3 -m pip install numpy
 
diff --git a/evaluation/src/containerized_eval.py b/evaluation/src/containerized_eval.py
index 533fb68274..d22f0cf895 100644
--- a/evaluation/src/containerized_eval.py
+++ b/evaluation/src/containerized_eval.py
@@ -29,6 +29,7 @@
 import eval_v
 import eval_lean
 import eval_dart
+import eval_el
 import tempfile
 
 
@@ -65,6 +66,7 @@
     "coq": (eval_v.eval_script, ".v"),
     "lean": (eval_lean.eval_script, ".lean"),
     "dart": (eval_dart.eval_script, ".dart"),
+    "el": (eval_el.eval_script, ".el"),
 }
 
 def eval_string_script(language, program):
diff --git a/evaluation/src/eval_el.py b/evaluation/src/eval_el.py
new file mode 100644
index 0000000000..5c14b3b352
--- /dev/null
+++ b/evaluation/src/eval_el.py
@@ -0,0 +1,33 @@
+"""
+Evaluates a generated Emacs Lisp program (.el).
+"""
+from pathlib import Path
+from safe_subprocess import run
+
+def eval_script(path: Path):
+
+    result = run([
+        "emacs", "-batch",
+        "-l", "ert",
+        "-l", str(path),
+        "-f", "ert-run-tests-batch-and-exit"
+    ])
+
+    if result.timeout:
+        status = "Timeout"
+    elif result.exit_code != 0:
+        status = "Exception"
+    elif "\nRan 1 tests, 1 results as expected, 0 unexpected" in result.stderr:
+        status = "OK"
+    else: # test failure
+        status = "Exception"
+
+    return {
+        "status": status,
+        "exit_code": result.exit_code,
+        "stdout": result.stdout,
+        "stderr": result.stderr,
+    }
+
+if __name__ == "__main__":
+    main()