From b1d608fd16e5cbc97ba4522520754b79fda43142 Mon Sep 17 00:00:00 2001
From: Dylan Pulver <dylanpulver@users.noreply.github.com>
Date: Fri, 21 Aug 2026 14:27:30 +0300
Subject: [PATCH] fix(langchain): Record reasoning tokens as metric

---
 .../langchain-reasoning-tokens-metric.md      |  5 ++
 .../langgraph-v1-latest.log-payloads.json     |  1 +
 .../langgraph-v1-latest.span-tree.json        |  1 +
 .../langgraph-v1-latest.span-tree.txt         |  1 +
 .../langgraph-v1.log-payloads.json            |  1 +
 .../__snapshots__/langgraph-v1.span-tree.json |  1 +
 .../__snapshots__/langgraph-v1.span-tree.txt  |  1 +
 ...ngchain-js-traces-v1-latest.span-tree.json |  6 ++
 ...angchain-js-traces-v1-latest.span-tree.txt |  6 ++
 ...wrap-langchain-js-traces-v1.span-tree.json |  6 ++
 .../wrap-langchain-js-traces-v1.span-tree.txt |  6 ++
 .../langchain/callback-handler.test.ts        | 62 +++++++++++++++++++
 js/src/wrappers/langchain/callback-handler.ts |  4 ++
 13 files changed, 101 insertions(+)
 create mode 100644 .changeset/langchain-reasoning-tokens-metric.md

diff --git a/.changeset/langchain-reasoning-tokens-metric.md b/.changeset/langchain-reasoning-tokens-metric.md
new file mode 100644
index 000000000..67b703aed
--- /dev/null
+++ b/.changeset/langchain-reasoning-tokens-metric.md
@@ -0,0 +1,5 @@
+---
+"braintrust": patch
+---
+
+fix(langchain): Record reasoning tokens as metric
diff --git a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.log-payloads.json b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.log-payloads.json
index 159e7c4ae..b64f5bec9 100644
--- a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.log-payloads.json
+++ b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.log-payloads.json
@@ -266,6 +266,7 @@
       "tags": []
     },
     "metrics": {
+      "completion_reasoning_tokens": "<number>",
       "completion_tokens": "<number>",
       "end": 0,
       "prompt_cached_tokens": "<number>",
diff --git a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.span-tree.json b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.span-tree.json
index 801834fae..b7429d6fb 100644
--- a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.span-tree.json
+++ b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.span-tree.json
@@ -113,6 +113,7 @@
                     "model": "gpt-4o-mini-2024-07-18"
                   },
                   "metric_keys": [
+                    "completion_reasoning_tokens",
                     "completion_tokens",
                     "prompt_cached_tokens",
                     "prompt_tokens",
diff --git a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.span-tree.txt b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.span-tree.txt
index 1c06fb02c..89da2a98e 100644
--- a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.span-tree.txt
+++ b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1-latest.span-tree.txt
@@ -113,6 +113,7 @@ span_tree:
         │         "model": "gpt-4o-mini-2024-07-18"
         │       }
         │       metric_keys: [
+        │         "completion_reasoning_tokens",
         │         "completion_tokens",
         │         "prompt_cached_tokens",
         │         "prompt_tokens",
diff --git a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.log-payloads.json b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.log-payloads.json
index 2b4b57071..c886ed5aa 100644
--- a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.log-payloads.json
+++ b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.log-payloads.json
@@ -265,6 +265,7 @@
       "tags": []
     },
     "metrics": {
+      "completion_reasoning_tokens": "<number>",
       "completion_tokens": "<number>",
       "end": 0,
       "prompt_cached_tokens": "<number>",
diff --git a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.span-tree.json b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.span-tree.json
index 801834fae..b7429d6fb 100644
--- a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.span-tree.json
+++ b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.span-tree.json
@@ -113,6 +113,7 @@
                     "model": "gpt-4o-mini-2024-07-18"
                   },
                   "metric_keys": [
+                    "completion_reasoning_tokens",
                     "completion_tokens",
                     "prompt_cached_tokens",
                     "prompt_tokens",
diff --git a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.span-tree.txt b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.span-tree.txt
index 1c06fb02c..89da2a98e 100644
--- a/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.span-tree.txt
+++ b/e2e/scenarios/langgraph-auto-instrumentation/__snapshots__/langgraph-v1.span-tree.txt
@@ -113,6 +113,7 @@ span_tree:
         │         "model": "gpt-4o-mini-2024-07-18"
         │       }
         │       metric_keys: [
+        │         "completion_reasoning_tokens",
         │         "completion_tokens",
         │         "prompt_cached_tokens",
         │         "prompt_tokens",
diff --git a/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1-latest.span-tree.json b/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1-latest.span-tree.json
index ad63c46a4..23cffa685 100644
--- a/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1-latest.span-tree.json
+++ b/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1-latest.span-tree.json
@@ -156,6 +156,7 @@
                 "tags": []
               },
               "metrics": {
+                "completion_reasoning_tokens": 0,
                 "completion_tokens": 2,
                 "prompt_cached_tokens": 0,
                 "prompt_tokens": 12,
@@ -422,6 +423,7 @@
                     ]
                   },
                   "metrics": {
+                    "completion_reasoning_tokens": 0,
                     "completion_tokens": 2,
                     "prompt_cached_tokens": 0,
                     "prompt_tokens": 18,
@@ -742,6 +744,7 @@
                 "tags": []
               },
               "metrics": {
+                "completion_reasoning_tokens": 0,
                 "completion_tokens": 6,
                 "prompt_cached_tokens": 0,
                 "prompt_tokens": 22,
@@ -974,6 +977,7 @@
                 "tags": []
               },
               "metrics": {
+                "completion_reasoning_tokens": 0,
                 "completion_tokens": 16,
                 "prompt_cached_tokens": 0,
                 "prompt_tokens": 72,
@@ -1233,6 +1237,7 @@
                 "tags": []
               },
               "metrics": {
+                "completion_reasoning_tokens": 0,
                 "completion_tokens": 21,
                 "prompt_cached_tokens": 0,
                 "prompt_tokens": 76,
@@ -1503,6 +1508,7 @@
                 "tags": []
               },
               "metrics": {
+                "completion_reasoning_tokens": 0,
                 "completion_tokens": 11,
                 "prompt_cached_tokens": 0,
                 "prompt_tokens": 106,
diff --git a/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1-latest.span-tree.txt b/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1-latest.span-tree.txt
index da37b45bb..ba4679a42 100644
--- a/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1-latest.span-tree.txt
+++ b/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1-latest.span-tree.txt
@@ -155,6 +155,7 @@ span_tree:
     │         "tags": []
     │       }
     │       metrics: {
+    │         "completion_reasoning_tokens": 0,
     │         "completion_tokens": 2,
     │         "prompt_cached_tokens": 0,
     │         "prompt_tokens": 12,
@@ -553,6 +554,7 @@ span_tree:
     │             ]
     │           }
     │           metrics: {
+    │             "completion_reasoning_tokens": 0,
     │             "completion_tokens": 2,
     │             "prompt_cached_tokens": 0,
     │             "prompt_tokens": 18,
@@ -716,6 +718,7 @@ span_tree:
     │         "tags": []
     │       }
     │       metrics: {
+    │         "completion_reasoning_tokens": 0,
     │         "completion_tokens": 6,
     │         "prompt_cached_tokens": 0,
     │         "prompt_tokens": 22,
@@ -940,6 +943,7 @@ span_tree:
     │         "tags": []
     │       }
     │       metrics: {
+    │         "completion_reasoning_tokens": 0,
     │         "completion_tokens": 16,
     │         "prompt_cached_tokens": 0,
     │         "prompt_tokens": 72,
@@ -1191,6 +1195,7 @@ span_tree:
         │     "tags": []
         │   }
         │   metrics: {
+        │     "completion_reasoning_tokens": 0,
         │     "completion_tokens": 21,
         │     "prompt_cached_tokens": 0,
         │     "prompt_tokens": 76,
@@ -1457,6 +1462,7 @@ span_tree:
               "tags": []
             }
             metrics: {
+              "completion_reasoning_tokens": 0,
               "completion_tokens": 11,
               "prompt_cached_tokens": 0,
               "prompt_tokens": 106,
diff --git a/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1.span-tree.json b/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1.span-tree.json
index 74a535155..9465136eb 100644
--- a/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1.span-tree.json
+++ b/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1.span-tree.json
@@ -156,6 +156,7 @@
                 "tags": []
               },
               "metrics": {
+                "completion_reasoning_tokens": 0,
                 "completion_tokens": 2,
                 "prompt_cached_tokens": 0,
                 "prompt_tokens": 12,
@@ -422,6 +423,7 @@
                     ]
                   },
                   "metrics": {
+                    "completion_reasoning_tokens": 0,
                     "completion_tokens": 2,
                     "prompt_cached_tokens": 0,
                     "prompt_tokens": 18,
@@ -742,6 +744,7 @@
                 "tags": []
               },
               "metrics": {
+                "completion_reasoning_tokens": 0,
                 "completion_tokens": 6,
                 "prompt_cached_tokens": 0,
                 "prompt_tokens": 22,
@@ -974,6 +977,7 @@
                 "tags": []
               },
               "metrics": {
+                "completion_reasoning_tokens": 0,
                 "completion_tokens": 16,
                 "prompt_cached_tokens": 0,
                 "prompt_tokens": 72,
@@ -1233,6 +1237,7 @@
                 "tags": []
               },
               "metrics": {
+                "completion_reasoning_tokens": 0,
                 "completion_tokens": 21,
                 "prompt_cached_tokens": 0,
                 "prompt_tokens": 76,
@@ -1503,6 +1508,7 @@
                 "tags": []
               },
               "metrics": {
+                "completion_reasoning_tokens": 0,
                 "completion_tokens": 11,
                 "prompt_cached_tokens": 0,
                 "prompt_tokens": 106,
diff --git a/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1.span-tree.txt b/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1.span-tree.txt
index cb09946d7..33d3688be 100644
--- a/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1.span-tree.txt
+++ b/e2e/scenarios/wrap-langchain-js-traces/__snapshots__/wrap-langchain-js-traces-v1.span-tree.txt
@@ -155,6 +155,7 @@ span_tree:
     │         "tags": []
     │       }
     │       metrics: {
+    │         "completion_reasoning_tokens": 0,
     │         "completion_tokens": 2,
     │         "prompt_cached_tokens": 0,
     │         "prompt_tokens": 12,
@@ -553,6 +554,7 @@ span_tree:
     │             ]
     │           }
     │           metrics: {
+    │             "completion_reasoning_tokens": 0,
     │             "completion_tokens": 2,
     │             "prompt_cached_tokens": 0,
     │             "prompt_tokens": 18,
@@ -716,6 +718,7 @@ span_tree:
     │         "tags": []
     │       }
     │       metrics: {
+    │         "completion_reasoning_tokens": 0,
     │         "completion_tokens": 6,
     │         "prompt_cached_tokens": 0,
     │         "prompt_tokens": 22,
@@ -940,6 +943,7 @@ span_tree:
     │         "tags": []
     │       }
     │       metrics: {
+    │         "completion_reasoning_tokens": 0,
     │         "completion_tokens": 16,
     │         "prompt_cached_tokens": 0,
     │         "prompt_tokens": 72,
@@ -1191,6 +1195,7 @@ span_tree:
         │     "tags": []
         │   }
         │   metrics: {
+        │     "completion_reasoning_tokens": 0,
         │     "completion_tokens": 21,
         │     "prompt_cached_tokens": 0,
         │     "prompt_tokens": 76,
@@ -1457,6 +1462,7 @@ span_tree:
               "tags": []
             }
             metrics: {
+              "completion_reasoning_tokens": 0,
               "completion_tokens": 11,
               "prompt_cached_tokens": 0,
               "prompt_tokens": 106,
diff --git a/js/src/wrappers/langchain/callback-handler.test.ts b/js/src/wrappers/langchain/callback-handler.test.ts
index e441fa1ad..1cf9682ee 100644
--- a/js/src/wrappers/langchain/callback-handler.test.ts
+++ b/js/src/wrappers/langchain/callback-handler.test.ts
@@ -128,4 +128,66 @@ describe("BraintrustLangChainCallbackHandler metrics", () => {
       tokens: 12,
     });
   });
+
+  it("preserves reasoning metrics from message usage metadata", async () => {
+    const { endLog } = await finishChatModelRun({
+      generations: [
+        [
+          {
+            message: {
+              usage_metadata: {
+                input_tokens: 10,
+                output_tokens: 20,
+                total_tokens: 30,
+                output_token_details: {
+                  reasoning: 16,
+                },
+              },
+            },
+          },
+        ],
+      ],
+    });
+
+    expect(endLog.metrics).toEqual({
+      prompt_tokens: 10,
+      completion_tokens: 20,
+      completion_reasoning_tokens: 16,
+      tokens: 30,
+    });
+  });
+
+  it("reports zero reasoning tokens rather than dropping the field", async () => {
+    const { endLog } = await finishChatModelRun({
+      generations: [
+        [
+          {
+            message: {
+              usage_metadata: {
+                input_tokens: 12,
+                output_tokens: 2,
+                total_tokens: 14,
+                input_token_details: {
+                  audio: 0,
+                  cache_read: 0,
+                },
+                output_token_details: {
+                  audio: 0,
+                  reasoning: 0,
+                },
+              },
+            },
+          },
+        ],
+      ],
+    });
+
+    expect(endLog.metrics).toEqual({
+      prompt_tokens: 12,
+      completion_tokens: 2,
+      prompt_cached_tokens: 0,
+      completion_reasoning_tokens: 0,
+      tokens: 14,
+    });
+  });
 });
diff --git a/js/src/wrappers/langchain/callback-handler.ts b/js/src/wrappers/langchain/callback-handler.ts
index a0143b338..9bf57bc51 100644
--- a/js/src/wrappers/langchain/callback-handler.ts
+++ b/js/src/wrappers/langchain/callback-handler.ts
@@ -523,6 +523,7 @@ function getMetricsFromResponse(
     }
 
     const inputTokenDetails = usageMetadata.input_token_details;
+    const outputTokenDetails = usageMetadata.output_token_details;
     return normalizeTokenMetrics({
       total_tokens: usageMetadata.total_tokens,
       prompt_tokens: usageMetadata.input_tokens,
@@ -533,6 +534,9 @@ function getMetricsFromResponse(
       prompt_cached_tokens: isRecord(inputTokenDetails)
         ? inputTokenDetails.cache_read
         : undefined,
+      completion_reasoning_tokens: isRecord(outputTokenDetails)
+        ? outputTokenDetails.reasoning
+        : undefined,
     });
   }