Kimi K3.1 Pro hits #8 on Code Arena WebDev, matching Claude Opus-5 at 70% lower cost
arena · x · 2026-10-24
Arena announced that Moonshot's Kimi K3.1 Pro scored 1692 on Code Arena: WebDev, jumping 9 places (#17→#8, +37 pts) from Kimi K3 and matching Claude Opus-5 (Max) at roughly 70% lower cost per task. Pricing is $2/$10 per million tokens with a median cost of $1.23 per task, 29% cheaper than K3; subcategory highlights include #7 in Brand & Marketing and #9 in Simulations.
Commentator mlangelopoulos calls it the #1 open-source model, placing it on par with Opus 5 from three months ago — implying China lags the frontier by 3 months — while noting Sonnet 5.5 + DeepSeek V4.1 Flash together dominate it on the performance-vs-cost Pareto frontier. "The race is so tight. The frontier never sleeps."
Related event: Kimi K3.1 Pro cracks Code Arena top 10 at a fraction of Claude's cost(4 posts)→
More from Models
- Terminal-Bench exposes NVIDIA Nemotron 3.5 near-zero score vs Kimi K3.1's 65.7% — JJitsev · 2026-10-24
- Bench shows GPT-5.6/6 Luna crush multi-agent coordination while DeepSeek V4 fails — JoshPurtell · 2026-10-24
- Grok Bot adds secure vault so agents can log in and pay without seeing your secrets — XFreeze · 2026-10-24
- Kimi K3.1 Pro reportedly first Chinese model to solve 20x20 nonograms — fairydreaming · 2026-10-24
- Kimi K3.1 vs GPT-6 Astra head-to-head on an identical agentic 3D task — GapNew4766 · 2026-10-24
- Kimi K3.1 Pro Launches: Scores 55 on Artificial Analysis, Open Weights Promised by Nov 2026 — Simon Willison · 2026-10-24