IFM's new K2-Horizon-MoVA-36B-A4B draws scrutiny: real deal or benchmaxxed?
edward-dev · reddit · 2026-09-04
Reddit users are asking how IFM's newly released MoE model K2-Horizon-MoVA-36B-A4B stacks up against similar-size competitors. It looks decent on benchmarks like terminal bench versus Qwen 3.6 35B A3B, but lacks SWE bench pro numbers, and little is known about the lab—leaving open whether it genuinely trades blows with models like tiel coder or is just benchmaxxed.
More from Models
- Martian says routing across 44 LLMs cuts errors 46% vs best single model on 16 benchmarks — Arindam_1729 · 2026-09-04
- Ex-OpenAI safety lead Miles Brundage: if your primary emotion on AI isn't concern, you're misreading it — Miles_Brundage · 2026-09-04
- Gary Marcus on GPT-6 Astra: symbolic world models are vindication, but no proof of AGI — GaryMarcus · 2026-09-04
- Ex-Google X Quant Guillaume Verdon Says OpenAI Is 'Kinda Back' — But Vibes, Not Benchmarks, Will Decide — beffjezos · 2026-09-04
- Mystery model "Astra" reportedly beats 5.6 Sol Pro on FrontierMath T4 — ctjlewis · 2026-09-04
- Inkling and Inkling-Small models get a free inference tier, and it's staying — simonguozirui · 2026-09-04