Unverified50% confidenceBenchmarkExact time
The first full run took 22.9 seconds, 12 parallel Jev calls completed in about 1.1 seconds, and the main performance bottleneck was Postgres queries of about 700ms each that also had to queue.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/13/2026
First Seen
Valid until: 10/27/2026
Sources
Related Entities
Related Claims
VerifiedJev completes classification tasks in only 70 to 500 milliseconds, which is 40 to 200 times faster than traditional LLMs.79% similarUnverifiedJev's bottleneck is a round-trip latency of over 200 milliseconds; at around 1,500 meters the game speed is already fast enough to render those responses meaningless.78% similarUnverifiedIn a game against Fable, Jev's chess strength was suppressed and it was trailing by 16 pieces at move 29, but Jev required only 2.6 seconds per move while Fable required 6 to 15 seconds per move, and Fable was ultimately adjudicated the loser for exceeding the time limit.76% similarUnverifiedAcross 14 tests covering approximately 145 skills, Jev found the correct skill in under 5 seconds on average, while using Opus 5 by default required approximately 30 seconds.76% similarUnverifiedOn a single CPU thread, Largha-150M-v1 takes about 1.64 seconds to process one case (one state plus five questions); with 8 threads, about 8–20 decisions/second.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1017058API
curl https://kongchang.com/api/v1/knowledge/claims/1017058MCP
get_claim(id=1017058)