# K2-Horizon-375B-A23B ![Organization](https://img.shields.io/badge/Org-IFM-555) ![Context](https://img.shields.io/badge/Context-512K%20tokens-1f6feb) ![Pricing](https://img.shields.io/badge/Price-Open%20weights-2da44e) ![License](https://img.shields.io/badge/License-Apache%202.0-d29922) ![Modality](https://img.shields.io/badge/Modality-Text-8957e5) ![Released](https://img.shields.io/badge/Released-Sep%202026-555) > **Quick answer:** K2-Horizon-375B-A23B is IFM's flagship open-weight Mixture-of-Experts model — 375B total parameters, 23B active per token, with a native 512K-token context window. On agentic tool-use, terminal, and long-horizon workflow benchmarks it matches or beats open-weight MoE models up to 2.6x its size and is competitive with closed frontier models. Open weights, Apache 2.0 license. ## At a Glance **Where K2-Horizon-375B-A23B leads** - Strong agentic tool use and terminal-use scores relative to its 23B active-parameter footprint - 512K native context window from midtraining onward - Fully open: weights, and (upcoming) training data/recipe/code **Where it lags** - Trails closed frontier models (GPT-5.6, Claude Sonnet 5) on GDPVal-AA Elo and BrowseComp - CritPt (frontier physics reasoning) is notably weaker than most peers in this comparison set **Best for:** teams wanting an open-weight model with frontier-adjacent agentic tool-use and terminal-use capability at a fraction of the active-parameter cost of larger MoE peers. ## What K2-Horizon-375B-A23B Is K2-Horizon-375B-A23B is the largest model in IFM's K2-Horizon family, a sparse Mixture-of-Experts architecture storing 375B parameters while activating only 23B per token. It is built for long-horizon agentic workflows — tool use, terminal interaction, and multi-step task completion — rather than pure static-benchmark chasing. The model ships with the final training checkpoint; intermediate checkpoints, training data, and training code are planned for future release, an unusually high transparency bar for a model at this scale. ## Specifications | Field | Value | |---|---| | Organization | IFM | | Parameters | 375B total / 23B active | | Context window | 524,288 tokens (512K) | | Architecture | Sparse Mixture-of-Experts (MoE) | | License | Apache 2.0 | | Release date | 2026-09 | | Modality | Text | ## Pricing | | Input (per 1M tokens) | Output (per 1M tokens) | |---|---|---| | Open weights | — | — | Open weights: free to download and self-host; cost is hosting/inference only. Validated serving recipes exist for vLLM and SGLang (8x H200, tensor-parallel 8). ## Context Window K2-Horizon-375B-A23B has a **524,288-token (512K) context window** — roughly 1,300+ pages of text in a single request. This is native from the midtraining stages onward, supporting long-document and long-horizon-agent workloads without external context-extension tricks. ## Public Benchmark Scores | Benchmark | Score | Source | Date | |---|---|---|---| | [GDPVal-AA v2](/benchmarks/artificial-analysis/gdpval-aa-v2) | 1,441 (Elo) | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [τ³ Banking](/benchmarks/tau-bench/tau3-banking) | 34.0 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [TerminalBench 2.1](/benchmarks/tbench/terminalbench-21) | 70.2 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [SciCode](/benchmarks/scicode-team/scicode) | 42.7 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [Humanity's Last Exam](/benchmarks/scale-ai/humanitys-last-exam) (no tools) | 32.0 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [GPQA Diamond](/benchmarks/rein-et-al/gpqa-diamond) | 87.3 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [CritPt](/benchmarks/critpt-team/critpt) | 8.6 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [AA-LCR](/benchmarks/artificial-analysis/aa-lcr) | 76.0 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [Toolathlon-Verified](/benchmarks/toolathlon/toolathlon-verified) | 65.3 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [AutomationBench](/benchmarks/zapier/automationbench) (Public subset) | 25.3 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [APEX-Agents](/benchmarks/apex/apex-agents) (pass@1, text subset) | 24.8 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [MCPMark-Verified](/benchmarks/mcpmark/mcpmark-verified) | 67.7 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [BrowseComp](/benchmarks/openai/browsecomp) | 72.8 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [WildClawBench](/benchmarks/wildclaw/wildclawbench) (English text subset) | 50.9 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | | [SWE Bench Pro](/benchmarks/scale-ai/swe-bench-pro) (strict, no internet) | 42.6 | [IFM model card](https://huggingface.co/IFM/K2-Horizon-375B-A23B) | 2026-09 | *AA-Omniscience Accuracy (23.0) and Non-Hallucination (74.7) were also reported but use a different metric structure than the platform's combined AA-Omniscience Index — pending mapping decision, not yet added. SWE-Atlas-QnA (48.4, strict) reported but pending exact-benchmark mapping confirmation. Scores above are reported by IFM and shown for context; not Benchgen measurements. GDPVal-AA is the Elo rating; other scores are % unless noted.* ## Use K2-Horizon-375B-A23B via API ```python from openai import OpenAI client = OpenAI(base_url="http://localhost:30000/v1", api_key="EMPTY") response = client.chat.completions.create( model="IFM/K2-Horizon-375B-A23B", messages=[{"role": "user", "content": "Explain the result step by step."}], temperature=1.0, top_p=0.95, extra_body={"chat_template_kwargs": {"reasoning_effort": "high"}}, ) print(response.choices[0].message.content) ``` ## Frequently Asked Questions

What is K2-Horizon-375B-A23B?

It's IFM's flagship open-weight Mixture-of-Experts model: 375B total parameters, 23B activated per token, with a 512K-token context window, built for agentic and long-horizon tasks.

What is K2-Horizon-375B-A23B's context window?

524,288 tokens (512K) — roughly 1,300+ pages of text in one request.

How much does K2-Horizon-375B-A23B cost?

Open weights — free to download; cost is hosting/inference only.

Is K2-Horizon-375B-A23B open source?

Yes, released under the Apache 2.0 license.

--- *Specs and scores sourced from IFM's official Hugging Face model card; third-party benchmark scores attributed inline. Last updated 2026-09-03.*

K2 Horizon 375B A23B — Evaluations

Evaluations