When was K2 Horizon MoVA 36B-A4B released?

K2 Horizon MoVA 36B-A4B was released on September 3, 2026 by ifm. This is the official K2 Horizon MoVA 36B-A4B release date tracked on LLM Stats.

How big is K2 Horizon MoVA 36B-A4B?

K2 Horizon MoVA 36B-A4B has 36 billion parameters. It was trained on 22.9 trillion tokens. It ships as an open-weight model, so you can download and run it on your own hardware.

Who created K2 Horizon MoVA 36B-A4B?

K2 Horizon MoVA 36B-A4B was created by ifm.

What is the license for K2 Horizon MoVA 36B-A4B?

K2 Horizon MoVA 36B-A4B is released under the Apache 2.0 license. This is an open-source / open-weight license that permits self-hosting.

Where is the K2 Horizon MoVA 36B-A4B paper or technical report?

K2 Horizon MoVA 36B-A4B has a paper or technical report available at https://ifm.ai/blog/k2/. Use that source for architecture, training, release and evaluation details.

What models should I compare K2 Horizon MoVA 36B-A4B against?

Common K2 Horizon MoVA 36B-A4B comparisons include K2 Horizon MoVA 36B-A4B vs GPT OSS 120B High, K2 Horizon MoVA 36B-A4B vs EXAONE 4.5 33B, K2 Horizon MoVA 36B-A4B vs LongCat-Flash-Thinking-2601. Compare them side by side for benchmark scores, pricing, context window, latency and provider availability.

K2 Horizon MoVA 36B-A4B Benchmarks, Pricing & Context Window

Name: K2 Horizon MoVA 36B-A4B
Author: ifm

K2 Horizon MoVA 36B-A4B: Benchmarks, Pricing & Context Window

K2 Horizon MoVA 36B-A4B is a language model from Institute of Foundation Models, released in September 2026.

K2 Horizon MoVA 36B-A4B is the sparse mid-size member of IFM's (MBZUAI) fully open K2 Horizon family: a Mixture-of-Experts reasoning model with 36B total and about 4B active parameters per token (48 layers, 100 routed experts with 8 active

K2 Horizon MoVA 36B-A4B benchmarks

Capability tiers

Standing within each category, adjusted for leaderboard depth.

Real tasks performance

High-confidence performance for K2 Horizon MoVA 36B-A4B across real-world prompt categories. Only 95% intervals at most 4 points wide are shown.

Performance by conversation depth

How K2 Horizon MoVA 36B-A4B holds up as conversations get longer.

Quality Tracker

K2 Horizon MoVA 36B-A4B Performance Across Datasets

Scores sourced from the model's scorecard, paper, or official blog posts

llm-stats.com - Fri Oct 30 2026

Notice missing or incorrect data?

K2 Horizon MoVA 36B-A4B model size

K2 Horizon MoVA 36B-A4B has 36 billion parameters and was trained on 22.9 trillion tokens. See how it compares to other models in the same parameter range.

ParametersTraining tokens

36BMoE

22.9Ttokens

636× tokens-to-params ratio

Large (30–80B)

36B

1B7B70B405B

Try now

huggle

K2 Horizon MoVA 36B-A4Bin Huggle

Make it with
K2 Horizon MoVA 36B-A4B.

K2 Horizon MoVA 36B-A4B latency

K2 Horizon MoVA 36B-A4B time to first token, sustained output throughput, and failed-request rate from live model usage over the trailing 7 days.

K2 Horizon MoVA 36B-A4B examples

Recent arena outputs from K2 Horizon MoVA 36B-A4B, picked from the highest-ranked matchups.

K2 Horizon MoVA 36B-A4B license

K2 Horizon MoVA 36B-A4B is released under the Apache 2.0 license, which permits commercial use, has 36.0B parameters.

License: Apache 2.0; Commercial use allowed
Parameters: 36.0B

Apache License 2.0 - allows commercial use

K2 Horizon MoVA 36B-A4B resources

Official sources for K2 Horizon MoVA 36B-A4B: official launch post, source repository, model weights.

K2 Horizon MoVA 36B-A4B vs other models

The most-compared alternatives to K2 Horizon MoVA 36B-A4B are GPT OSS 120B High, EXAONE 4.5 33B, LongCat-Flash-Thinking-2601. Open any pair side-by-side for benchmarks, pricing, context, and latency.