Expired50% confidenceBenchmarkExact time
On agent turns, with 1 concurrent agent the median is 186 tok/s and p5 is 134 tok/s; with 4 it is 59 tok/s median and 52 tok/s p5; with 6 it is 40 tok/s median and 34 tok/s p5. The article sets 40 tok/s per person as the standard, and four concurrent streams with 10% headroom is this machine's capacity.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/16/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/30/2026(expired)
Sources
Related Entities
Related Claims
Unverified一个Agent的总Token消耗可能达到50K-100K，而同样的对话在Chat模式下可能只需要5K68% similarUnverifiedWhen running IQ3_XXS quantized models with llama.cpp, the output speed is only 15 tok/s with prompt processing at 100–120 tok/s, while a self-developed engine brings the same quantization up to approximately 65 tok/s output and 430 tok/s prompt processing, representing close to a 4x performance improvement.65% similarVerifiedA provider's rate limits typically take effect simultaneously across three dimensions: requests per minute (RPM), tokens per minute (TPM), and tokens per day (TPD).63% similarUnverifiedIn Omega Agent's domain modeling, Budget limits max turns (e.g., 100 turns), max tokens (e.g., 5 million), and max seconds (e.g., 60 seconds for a single tool call).63% similarUnverifiedUnder a 4M TPM premise, long Agent turns with a 400K context can trigger at most 10 requests per minute (hitting the TPM cap), and full-context requests only 4 per minute.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1044822API
curl https://kongchang.com/api/v1/knowledge/claims/1044822MCP
get_claim(id=1044822)