Kimi K3.1 Pro tops SWE Marathon and agent benchmarks, fueling open-source 'last time they catch up' debate
cephaloform · x · 2026-10-24
A multi-benchmark chart shared on X shows Kimi K3.1 Pro topping SWE Marathon (51.7), Kimi Code Bench 3.0 Long-horizon (61.2), Agents' Last Exam (43.2), OfficeQA Pro v2, Humanity's Last Exam with tools (78.0), and GDP.pdf — trading wins with Opus 5.5, GPT-6 Astra, GLM-5.3, and DS-V4.1. Frontier open-weights models are now effectively at parity with closed leaders.
The poster's jab: every time open models catch up, skeptics move the goalposts to "okay but this is the last time they caught up" — a prediction that keeps failing.
More from Models
- Community claims Kimi k3.1 is the most 'benchmaxxed' model ever, alleging distillation — imjustnewatai · 2026-10-24
- Leaker: Grok 4.8 arriving next week — mark_k · 2026-10-24
- Labs are far ahead of what they release — and the gap may be widening, speculates X user — MoonL88537 · 2026-10-24
- API Pricing Cut from $15 to $10 per Million Tokens — cedric_chee · 2026-10-24
- MiniMax's 430B model beats 3T-scale rivals on AA knowledge benchmark — teortaxesTex · 2026-10-24
- Hand-crafted regex pretokenizers quietly do heavy lifting in every frontier LLM — YouJiacheng · 2026-10-24