--- base_model: IFM/K2-Horizon-7B library_name: peft license: apache-2.0 pipeline_tag: text-generation tags: - peft - lora - diffusion-language-model - conditional-lora - uno ---

K2-Horizon-7B-Uno

**Paper:** [Unlocking Lossless Speedups in LLMs via Discrete Diffusion](https://huggingface.co/papers/2609.04010) **Project Page:** [https://s-sahoo.github.io/uno/](https://s-sahoo.github.io/uno/) **Code:** [https://github.com/ifm-ai/uno](https://github.com/ifm-ai/uno) We present `K2-Horizon-7B-Uno`, a diffusion-augmented LLM with two pathways in a unified architecture: - The AR pathway uses the AR weights of the [K2-Horizon-7B](https://huggingface.co/IFM/K2-Horizon-7B) model. - The diffusion pathway augments these weights with LoRA-based diffusion adapters. ![results_overview](./results_overview.png) This repository contains the adapter only. The base-model weights are hosted separately in [**IFM/K2-Horizon-7B**](https://huggingface.co/IFM/K2-Horizon-7B). ## Code Selected evaluation scripts are available in [**scripts/k2_horizon**](https://github.com/ifm-ai/uno/tree/main/scripts/k2_horizon). The full evaluation suite will be released soon. ## Evaluation results The main number is the benchmark score and the subscript is TPF. Bold marks the best score in each row. `--` and `N/A` denote unavailable results.

	Ours	Reference models
Benchmark	Uno _8B	Nemotron-Diff _14B	Mercury 2 _{size N/A}	Diff-Gemma _26B-A4B
Agentic Tool Use
τ³ Banking	25.8_--	--	9_N/A	--
τ² Telecom	90.1_2.46	14.3_4.8	71_N/A	68.1_18.8
τ² Retail	67.1_3.00	5.6_3.1	--	65.5_23.7
Terminal-Bench v2.1	39.1_2.65	4.5_7.5	27_N/A	14.7_14.1
Agentic Coding
SWE-bench Verified	70.1_3.08	0.8_1.5	--	18.7_5.6
Long-Context Reasoning
AA-LCR	68.0_2.85	7.3_1.1	36_N/A	19.7_10.8
Science and Knowledge
Humanity's Last Exam	18.6_2.47	2.6_7.2	16_N/A	9.2_14.1
GPQA-Diamond	77.1_2.71	40.4_7.6	77_N/A	70.7_11.9
AA-Omniscience	14.3_2.23	11.0_11.1	20_N/A	17.7_9.9
Math
GSM8K	95.4_2.88	93.1_6.1	--	95.1_28.9
MATH500	98.9_2.86	89.2_5.6	--	92.4_24.1
AIME-24	93.0_2.90	56.7_4.9	--	73.7_16.9
AIME-25	90.7_2.96	40.0_4.5	--	74.3_18.9
AIME-26	86.3_2.84	46.7_4.8	--	70.7_17.8
Coding
MBPP	84.1_2.44	73.8_5.3	--	80.1_15.5
HumanEval	95.2_2.66	84.8_7.5	--	95.1_28.2
Efficiency Summary
Average TPF	2.71	5.41	N/A	17.56
System Throughput	5,255	2,794	1,197	1,136
Per-request Throughput	405	290	769	836

## Citation If you find this model useful, please cite: ```bibtex @misc{k2_horizon_7b_uno, title = {K2-Horizon-7B-Uno}, author = {Institute of Foundation Models}, year = {2026}, howpublished = {\url{https://huggingface.co/IFM/K2-Horizon-7B-Uno}}, } ```