All AI models · LLM benchmarks · Methodology

K2-Horizon-7B

IFM's open-weight 7B model for coding, reasoning, and chat. Heabsy serves an FP8 deployment on a dedicated RTX 5090 with speculative decoding in the United States, with a 131K context window and an 8K output limit. Zero data retention is not verified for this deployment.

Model facts

Context window
131K tokens
Maximum output
8K tokens
Cheapest paid input
$0.05 per 1M tokens
Cheapest paid output
$0.2 per 1M tokens
Open weights
yes
Providers
1

Capabilities and modalities

reasoning, tool calling, open weights, text.

Local hardware estimate

At 4K context: Q4_K_M 6.5 GiB, Q8_0 10.7 GiB, F16 18.5 GiB working memory.

Parameters
9.0 billion
Architecture
k2horizon
Model license
apache-2.0

Planning estimate adapted from the MIT-licensed whichllm estimator using metadata from Hugging Face; not a vendor minimum requirement.

Observed price history

  • 2026-09-10: $0.05 input / $0.2 output per 1M tokens

API providers