From aff84a5d9e4d16e89530ab0b2abaa0717fd7f80a Mon Sep 17 00:00:00 2001 From: Mark Backman Date: Tue, 19 May 2026 11:05:15 -0400 Subject: [PATCH 1/3] Add P99 latency for Smallest AI STT --- src/pipecat/services/stt_latency.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/pipecat/services/stt_latency.py b/src/pipecat/services/stt_latency.py index 7d285e353..b7411f011 100644 --- a/src/pipecat/services/stt_latency.py +++ b/src/pipecat/services/stt_latency.py @@ -47,6 +47,7 @@ GROQ_TTFS_P99: float = 1.54 OPENAI_TTFS_P99: float = 2.01 OPENAI_REALTIME_TTFS_P99: float = 1.66 SARVAM_TTFS_P99: float = 1.17 +SMALLEST_TTFS_P99: float = 1.59 SONIOX_TTFS_P99: float = 0.35 SPEECHMATICS_TTFS_P99: float = 0.74 @@ -56,5 +57,4 @@ WHISPER_TTFS_P99: float = DEFAULT_TTFS_P99 # No benchmark available yet; using conservative default MISTRAL_TTFS_P99: float = DEFAULT_TTFS_P99 -SMALLEST_TTFS_P99: float = DEFAULT_TTFS_P99 XAI_TTFS_P99: float = DEFAULT_TTFS_P99 From 33e5d1f89b7856bf863da2ed4e764682c823f77a Mon Sep 17 00:00:00 2001 From: Mark Backman Date: Tue, 19 May 2026 11:23:49 -0400 Subject: [PATCH 2/3] Add changelog for PR #4522 --- changelog/4522.changed.md | 1 + 1 file changed, 1 insertion(+) create mode 100644 changelog/4522.changed.md diff --git a/changelog/4522.changed.md b/changelog/4522.changed.md new file mode 100644 index 000000000..37156adc7 --- /dev/null +++ b/changelog/4522.changed.md @@ -0,0 +1 @@ +- Updated the default p99 TTFS latency values for Smallest AI, Mistral, and XAI STT so turn stop timing uses measured values instead of the conservative fallback. From 38407e091de9748c2bb07da10c22a705c4fd6d61 Mon Sep 17 00:00:00 2001 From: Mark Backman Date: Tue, 19 May 2026 18:28:07 -0400 Subject: [PATCH 3/3] Add p99 values for Mistral and XAI --- src/pipecat/services/stt_latency.py | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/src/pipecat/services/stt_latency.py b/src/pipecat/services/stt_latency.py index b7411f011..f3fa235d5 100644 --- a/src/pipecat/services/stt_latency.py +++ b/src/pipecat/services/stt_latency.py @@ -44,17 +44,15 @@ GLADIA_TTFS_P99: float = 1.49 GOOGLE_TTFS_P99: float = 1.57 GRADIUM_TTFS_P99: float = 1.61 GROQ_TTFS_P99: float = 1.54 +MISTRAL_TTFS_P99: float = 1.89 OPENAI_TTFS_P99: float = 2.01 OPENAI_REALTIME_TTFS_P99: float = 1.66 SARVAM_TTFS_P99: float = 1.17 SMALLEST_TTFS_P99: float = 1.59 SONIOX_TTFS_P99: float = 0.35 SPEECHMATICS_TTFS_P99: float = 0.74 +XAI_TTFS_P99: float = 2.14 # These services run locally and should be replaced with measured values NVIDIA_TTFS_P99: float = DEFAULT_TTFS_P99 WHISPER_TTFS_P99: float = DEFAULT_TTFS_P99 - -# No benchmark available yet; using conservative default -MISTRAL_TTFS_P99: float = DEFAULT_TTFS_P99 -XAI_TTFS_P99: float = DEFAULT_TTFS_P99