Physix FrontierPhysix Frontier · News Briefing Card (Hacker News · Oct 24, 2026)
Kimi K3.1 Pro Cuts Open-Closed Model Gap to 31 Days
KEY FACTS
- Kimi released K3.1 Pro on October 23, scoring 55.1 on the Artificial Analysis Intelligence Index.
- That score is 2.5 points below the strongest closed model, Claude Opus 5.5, and 8.8 points above the previous best open model.
- Kimi says the model weights will be freely downloadable by November 4, having previously been available only through its own API.
- Claude Opus 5.5 reached that score 31 days earlier than Kimi, shrinking the open-source lag from 104 days.
- Chinese labs have held the best open-model title for 765 consecutive days, with six labs including Alibaba and DeepSeek taking turns at the top.
KEY DATA
55.1K3.1 Pro Intelligence Index score
2.5 pointsGap to Claude Opus 5.5
8.8 pointsGap to previous best open model
input $2/million tokens, output $10/million tokensAPI price
PHYSIX OBSERVATION
The gap between the strongest open and closed models has compressed from three months to one, a qualitative signal built from quantitative accumulation. Kimi delivers near-parity performance at the same price as Sonnet 5.5, and once the weights are open, enterprises and developers can self-host, shifting bargaining power toward buyers. But with six Chinese labs taking turns at the top and the leader holding the crown for only about 30 days on average, the lead window is extremely short. The real moat is not a one-time leaderboard win, but sustained iteration and ecosystem lock-in.
Source: Hacker News report