When was K2 Horizon 32B released?

K2 Horizon 32B was released on September 3, 2026 by ifm. This is the official K2 Horizon 32B release date tracked on LLM Stats.

How big is K2 Horizon 32B?

K2 Horizon 32B has 32 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 32B?

K2 Horizon 32B was created by ifm.

What is the license for K2 Horizon 32B?

K2 Horizon 32B 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 32B paper or technical report?

K2 Horizon 32B 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 32B against?

Common K2 Horizon 32B comparisons include K2 Horizon 32B vs Gemma 4 31B, K2 Horizon 32B vs LongCat-Flash-Thinking-2601, K2 Horizon 32B vs MAI-Code-1.1-Flash. Compare them side by side for benchmark scores, pricing, context window, latency and provider availability.

K2 Horizon 32B Benchmarks, Pricing & Context Window

Name: K2 Horizon 32B
Author: ifm

K2 Horizon 32B: Benchmarks, Pricing & Context Window

K2 Horizon 32B is a language model from Institute of Foundation Models, released in September 2026.

K2 Horizon 32B is the large dense member of IFM's (MBZUAI) fully open K2 Horizon family: a 32B-class decoder-only reasoning model (64 layers, 250K vocabulary; about 34.8B parameters in the released safetensors) pretrained on 22.9T tokens,

K2 Horizon 32B benchmarks

Capability tiers

Standing within each category, adjusted for leaderboard depth.

Real tasks performance

High-confidence performance for K2 Horizon 32B across real-world prompt categories. Only 95% intervals at most 4 points wide are shown.

Performance by conversation depth

How K2 Horizon 32B holds up as conversations get longer.

Quality Tracker

K2 Horizon 32B 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 32B model size

K2 Horizon 32B has 32 billion parameters and was trained on 22.9 trillion tokens. See how it compares to other models in the same parameter range.

ParametersTraining tokens

32B

22.9Ttokens

716× tokens-to-params ratio

Large (30–80B)

32B

1B7B70B405B

Try now

huggle

K2 Horizon 32Bin Huggle

Make it with
K2 Horizon 32B.

K2 Horizon 32B latency

K2 Horizon 32B time to first token, sustained output throughput, and failed-request rate from live model usage over the trailing 7 days.

K2 Horizon 32B examples

Recent arena outputs from K2 Horizon 32B, picked from the highest-ranked matchups.

K2 Horizon 32B license

K2 Horizon 32B is released under the Apache 2.0 license, which permits commercial use, has 32.0B parameters.

License: Apache 2.0; Commercial use allowed
Parameters: 32.0B

Apache License 2.0 - allows commercial use

K2 Horizon 32B resources

Official sources for K2 Horizon 32B: official launch post, source repository, model weights.

K2 Horizon 32B vs other models

The most-compared alternatives to K2 Horizon 32B are Gemma 4 31B, LongCat-Flash-Thinking-2601, MAI-Code-1.1-Flash. Open any pair side-by-side for benchmarks, pricing, context, and latency.