IFM/K2-Horizon-0.9B

Overview

K2-Horizon-0.9B is a compact dense model in the IFM K2-Horizon family, with 1.08B stored parameters including embeddings and a 131,072-token context window.

K2-Horizon is an open-weight IFM model family built for transparent foundation-model research, staged checkpoint analysis, and practical deployment. The series spans compact dense models such as K2-Horizon-0.9B for local experimentation, dense mid-size and large models for high-quality research workloads, and mixture-of-expert (MoE) models for higher-capacity serving.

Launch Commands

K2-Horizon-0.9B ships YaRN RoPE scaling in its config.json (factor 16 over an 8,192-token original window, rope_theta 1,000,000, beta_fast 128, beta_slow 4), so the full 131,072-token context needs no --hf-overrides. A top-level rope_parameters override replaces that whole block rather than merging into it, so a partial one silently drops rope_theta and the YaRN betas.

Recommended: With YaRN

vllm serve IFM/K2-Horizon-0.9B \
  --trust-remote-code \
  --chat-template-content-format string \
  --dtype bfloat16 \
  --max-model-len 131072 \
  --gpu-memory-utilization 0.85 \
  --reasoning-parser k2_horizon \
  --enable-auto-tool-choice \
  --tool-call-parser k2_horizon

This configuration enables the full 131,072-token context length and is the recommended setting for K2-Horizon.

Shorter context

To cap memory at the original 8,192-token window, lower --max-model-len. YaRN stays on because it comes from the checkpoint config:

vllm serve IFM/K2-Horizon-0.9B \
  --trust-remote-code \
  --chat-template-content-format string \
  --dtype bfloat16 \
  --max-model-len 8192 \
  --gpu-memory-utilization 0.85 \
  --reasoning-parser k2_horizon \
  --enable-auto-tool-choice \
  --tool-call-parser k2_horizon

Client usage

from openai import OpenAI

client = OpenAI(
    api_key="EMPTY", base_url="http://localhost:8000/v1", timeout=3600
)

resp = client.chat.completions.create(
    model="IFM/K2-Horizon-0.9B",
    messages=[{"role": "user", "content": "Give me three primes above 100."}],
    temperature=1.0, top_p=0.95, max_tokens=2048,
)
print(resp.choices[0].message.content)

Thinking modes

The model supports selectable thinking through chat_template_kwargs, per request or server-wide via --default-chat-template-kwargs:

• {"reasoning_effort": "high"} — full thinking (default).
• {"reasoning_effort": "medium"} — faster thinking.
• {"reasoning_effort": "low"} — fastest thinking.

References

• IFM/K2-Horizon-0.9B on Hugging Face
• K2 Horizon Collection