Unverified50% confidenceBenchmarkExact time
Under load, the P99 TTFT of Agent sessions drops by 9.5 times; when the frontend is saturated by long requests, this figure reaches 23 times.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/15/2026
First Seen
Valid until: 1/13/2027
Sources
Related Entities
Related Claims
UnverifiedWhen a user hits the API rate limit, the API typically returns HTTP status code 429 (Too Many Requests).62% similarUnverifiedOverall P90 tail latency decreased by 22%.61% similarUnverifiedWhen running Whisper ASR, a TTS voice model, and nomic-embed-text simultaneously on an RTX 3090 with 24 GB of VRAM, the three resident models occupied 17.8 GB (72%) of VRAM, peaked at 23.4 GB over the course of a week, and never spilled over into system memory.59% similarUnverifiedThrough request-owned speculative KV, in a DeepSeek-V4-Pro 1P1D, MI355X, concurrency-256 scenario, the average TTFT was reduced from 19.4 seconds to 15.1 seconds.59% similarUnverifiedUsing a network debugging assistant directly for TCP speed testing yields only approximately 31 Mbps, because the debugging assistant does not reply with ACKs fast enough.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1038166API
curl https://kongchang.com/api/v1/knowledge/claims/1038166MCP
get_claim(id=1038166)