// AI Platform Drift

langchain-fireworks 1.7.2 fixes streamed token usage reporting

LangChain releases · published 2026-11-11 06:23 UTC
WORTH KNOWING VERIFY_ONLY API
WHAT CHANGEDlangchain-fireworks 1.7.2 was released on 2026-11-04. The package now reports streamed token usage once instead of multiple times, includes served model metadata in response objects, promotes reranking out of beta, and relaxes third-party dependency upper bounds.
WHY IT MATTERSIf you parse token usage from streamed Fireworks responses, you were receiving duplicate counts. The fix changes what your usage tracking sees. The served model metadata addition means response objects now contain which model actually handled the request, which may differ from what you requested if Fireworks routed to an equivalent.
WHAT TO DOUpgrade to 1.7.2 and verify your token usage accounting. If you sum usage from stream chunks, check that totals now match non-streamed calls. If you log or bill by model, confirm your code handles the new served_model field in response metadata.
Model-reported confidence 75% (not measured accuracy) Primary source ↗ written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+## langchain-fireworks==1.7.2
+https://github.com/langchain-ai/langchain/releases/tag/langchain-fireworks%3D%3D1.7.2
+2026-11-04T20:47:52Z
+
+Changes since langchain-fireworks==1.7.1
+fix(fireworks): report streamed token usage once (#41693)
+release(fireworks): 1.7.2 (#41690)
+fix(fireworks): report the served model in response metadata (#41686)
+chore(model-profiles): refresh model profile data (#41669)
+fix(fireworks): promote reranking out of beta (#41551)
+chore(model-profiles): refresh model profile data (#41432)
+chore(infra): relax third-party dependency upper bounds (#41407)
+chore(model-profiles): refresh model profile data (#41354)
+chore(deps): bump certifi from 2025.8.3 to 2026.10.19 in /libs/partners/fireworks (#41387)
+chore(deps): bump aiohttp from 3.14.4 to 3.14.5 in /libs/partners/fireworks (#41359)
+chore(deps): bump orjson from 3.12.0 to 3.13.0 in /libs/partners/fireworks (#41340)
+chore(deps): bump aiohttp from 3.14.3 to 3.14.4 in /libs/partners/fireworks (#41314)
+chore(model-profiles): refresh model profile data (#41298)
+chore(model-profiles): refresh model profile data (#41265)
+fix: fix dependencies in some partner packages (#41266)
+test(fireworks): stabilize forced tool-call extraction (#41139)
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
triage verdictMATERIAL at 0.72, needed 0.5
triage reasoningReranking promotion and token usage reporting changes affect API behavior and feature availability.
signalsreranking feature promoted out of beta · token usage reporting fix for streamed responses · served model metadata now included in response · dependency upper bounds relaxed
suppression—
triage modelanthropic/claude-haiku-4.5 ($0.00000)
writer modelanthropic/claude-sonnet-4.5 ($0.00000)