K2-Horizon-MoVA-36B-A4B-FP8

[!NOTE] This repository contains an FP8-quantized version of IFM/K2-Horizon-MoVA-36B-A4B.

Only the routed-expert linear layers in transformer layers 3–47 are quantized to FP8:

• Weights: static FP8 E4M3FN, one FP32 scale per 128*128 block.
• Activations: dynamic FP8, one scale per 1*128 group along the input-channel dim.

All other source-model tensors (attention, shared experts, routers, the first 3 dense layers, MoVA value experts, norms, embeddings, and lm_head) are kept in BF16.

The FP8 model performs closely in line with the original BF16 model on our evaluations, while reducing memory footprint and enabling faster inference on FP8-capable hardware.

Serving note: the routed experts' intermediate size (768) is splittable into whole 128-wide quantization blocks at TP=1 and TP=2, so these tensor-parallel sizes are recommended.

K2-Horizon-MoVA-36B-A4B is the sparse member of the K2-Horizon family: a Mixture-of-Experts model with Mixture-of-Values attention (MoVA) that stores 36B parameters and runs 4B per token. We have released the final checkpoint; intermediate checkpoints, along with the data and the training code, will be released.

K2-Horizon-MoVA-36B-A4B Highlights

• Frontier-class results at 4B active parameters. On agentic and reasoning benchmarks it outscores open weight dense (approximately 30B model size) and MoE models up to 15× its size; and also performs competitively against closed frontier models (see Benchmark Results).
• 512K context. Native 524,288-token context from the midtraining stages onward.
• Intermediate checkpoints. Intermediate checkpoints will be released so capability changes can be studied across training rather than at a single checkpoint.
• Fully open. Training data/recipe and the training code will be made public.

Benchmark Results

Quickstart

Serving

vLLM, recipe at recipes.vllm.ai/IFM:

vllm serve IFM/K2-Horizon-MoVA-36B-A4B-FP8 \
  --revision main \
  --tensor-parallel-size 2 \
  --enable-expert-parallel \
  --trust-remote-code \
  --max-model-len 131072 \
  --reasoning-parser k2_horizon \
  --tool-call-parser k2_horizon \
  --enable-auto-tool-choice

SGLang, adapted from the BF16 recipe validated on 2× H200 in the SGLang K2 Horizon cookbook:

python3 -m sglang.launch_server \
  --model-path IFM/K2-Horizon-MoVA-36B-A4B-FP8 \
  --revision main \
  --tp 2 \
  --ep 2 \
  --attention-backend fa3 \
  --json-model-override-args '{"xllm_source_router_gemm_partitions":2}' \
  --reasoning-parser k2_horizon \
  --tool-call-parser k2_horizon \
  --host 0.0.0.0 --port 30000

API Usage

[!Tip] Recommended settings: reasoning_effort="high", temperature=1.0, top_p=0.95. Reasoning depth is selected per request through chat_template_kwargs. Thinking is returned in reasoning_content and the answer in content.

from openai import OpenAI

client = OpenAI(base_url="http://localhost:30000/v1", api_key="EMPTY")
response = client.chat.completions.create(
    model="IFM/K2-Horizon-MoVA-36B-A4B-FP8",
    messages=[{"role": "user", "content": "Explain the result step by step."}],
    temperature=1.0,
    top_p=0.95,
    max_tokens=32768,
    extra_body={"chat_template_kwargs": {"reasoning_effort": "high", "tool_call_format": "xml"}},
)
message = response.choices[0].message
print("Reasoning:", getattr(message, "reasoning_content", None))
print("Answer:", message.content)

Our model supports multiple tool calls formats, which can be changed with chat_template_kwargs. The supported values are json, xml, and xml_typed . The default is xml. Keep --tool-call-parser k2_horizon enabled to parse the selected format.

Transformers

Validated with Transformers 5.15.0, PyTorch 2.13.0, Safetensors 0.8.0.

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "IFM/K2-Horizon-MoVA-36B-A4B-FP8"
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_id, device_map="auto", dtype="auto", low_cpu_mem_usage=True, trust_remote_code=True
)

inputs = tokenizer("Explain why long-context evaluation is difficult.", return_tensors="pt").to(model.device)
inputs.pop("token_type_ids", None)
outputs = model.generate(**inputs, max_new_tokens=32768, temperature=1.0, top_p=0.95, do_sample=True)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Best Practices

1. Reasoning effort: always high. All reported results use high reasoning effort. Pass {"chat_template_kwargs": {"reasoning_effort": "high"}} on every request.
2. Sampling parameters. temperature=1.0, top_p=0.95.
3. Serving. Use the SGLang command above: TP=2, FlashAttention-3, and the xllm_source_router_gemm_partitions override, which preserves the checkpoint's router numerics. Precision is read from the checkpoint, so no --dtype flag is needed. Full recipes for every K2-Horizon size, with measured H200 latency and throughput, are in the SGLang cookbook and the vLLM recipe.
4. Parsers. Enable the k2_horizon reasoning parser for chat, and add the k2_horizon tool-call parser for agent use. Leave both off for plain completion-style generation.

Citation

@misc{k2horizon2026,
  title  = {Introducing K2 Horizon: Frontier Performance, Radically Open},
  author = {{IFM Team}},
  year   = {2026},
  url    = {https://ifm.ai/blog/k2/},
}