Unverified50% confidenceBenchmarkExact time
Long document processing (the prefill phase) can reach a speed of 140 tokens per second.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/9/2026
First Seen
Valid until: 2/7/2027
Sources
Running Local LLMs on an Intel MacBook: A 400-Line Python Management Panel Put to the Test
bilibili人工智能打老虎11/9/2026
Related Entities
Related Claims
VerifiedAt a normal speaking pace, one can output 150–200 characters per minute, far exceeding typing speed.70% similarUnverifiedAn agent task requiring 20 LLM calls with 200 tokens generated each time takes approximately 133 seconds to complete at 30 TPS, but only 13 seconds at 300 TPS.68% similarUnverified170 tokens/s is the combined speed across all concurrent tasks, not the individual speed of a single agent.68% similarUnverifiedThe prefill phase uses small expert-parallel groups of 2-8 GPUs; when the batch increases from about 4,000 tokens to 16,000 tokens, throughput rises from about 60,000 tokens/s to 82,000 tokens/s, and per-step latency increases from about 70 milliseconds to 200 milliseconds.68% similarUnverifiedIt is generally considered that single-concurrency output above 100 tokens/second is sufficient to achieve a smooth typewriter effect, and at approximately 200 tokens/second, generating a 2,000-character long text takes only about 10 seconds.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1152902API
curl https://kongchang.com/api/v1/knowledge/claims/1152902MCP
get_claim(id=1152902)