explainx.ai350K+ learn from us
TrendingAI News TodayPathwaysSkills
Pricing
explainx.ai

Upskill in AI — 16 free pathways, live workshops & bootcamps, and 50+ courses from practitioners. Plus the skills, tools, and MCP servers to practice on.

follow us

follow on google

Add explainx.ai as a preferred source

corporate training

support@explainx.ai

get started

Find your pathTake Free Evaluation

community

Join the community

learn

mind: share how you thinkpathways — start freeworkshopsbootcampscoursescompare Explainxcertificationsmock testsexplainx universitycorporate traininglearn skills & mcp

discover

skillsmcp serversexplainx mcpestacktoolsmdx readeragentsllmsdesignsdictionarypeopleagi trackerfelony benchranks

company

aboutvisionmissionteaminstructorsteach on explainxpartnershipscommunityhackathonscareers

content

daily AI newsstate of AI — live resultsblogreleasespromptsgeneratorsresource libraryfor LLMsexplainx.ai kids

solutions

all solutionsdeveloper upskillingmarketing upskillingproduct manager upskillingleadership upskilling

newsletter · weekly

Get AI news, tools, and insights in your inbox.

supportcontactprivacytermsdata rightshow we create contentsubmission guidelines

© 2026 AISOLO Technologies Pvt Ltd

explainx.ai

On this page

  • TL;DR
  • What changed versus older K2-Horizon mentions
  • Modeling ideas builders will actually touch
  • The openness package: what to clone first
  • Performance claims and honest benchmarking
  • How to try K2-Horizon this week
  • Limits and verification gaps
  • What developers are asking
  • Related reading
  • Model-by-model: who each Horizon size is for
  • Synthetic data at 10T tokens: why builders should read the pipelines
  • xLLM and RL360: infrastructure you can diff against your stack
  • Risk checklist before production
  • What people search after a Horizon announcement
← Back to blog

explainx / blog

K2-Horizon: IFM Six-Model Open Fleet From 0.9B to 375B MoE

Open Source, Open Weights, MoE, IFM, Model Releases

Part of AI Tools and Apps

IFM released K2-Horizon: six open models, MoVA attention, Uno speedups, 22T-token shared pretraining, and LLM360-level training artifacts at ifm.ai/k2.

Nov 17, 2026·9 min read·Yash Thakker
add explainx.ai
go deep
K2-Horizon: IFM Six-Model Open Fleet From 0.9B to 375B MoE

IFM published K2-Horizon on November 15–17, 2026 as a six-model open fleet scaling from a 0.9B dense checkpoint to a 375B-A23B mixture-of-experts, with training data, intermediate checkpoints, and infrastructure code—not just inference weights.** The Hugging Face paper page summarizes modeling contributions (MoVA, softplus attention gating, Uno diffusion adapters) and states four models trained on a shared 22 trillion-token corpus.

For builders comparing open stacks this week against Qwen Code agents or Tencent Hy4 API drops, K2-Horizon’s pitch is radical reproducibility: IFM says it releases the pipelines that produced roughly 10T synthetic pretraining tokens, not only the merged checkpoints.

Daily briefFrom Yash · 350K+ students taught

Catch up on AI

One AI story explained in plain English, plus one thing to try. Every day, free. Unsubscribe anytime.

TL;DR

table · 2 cols
Model (IFM)Role in fleet
375B-A23B MoEEnterprise flagship for coding and agentic work
36B-A4B MoECost-efficient agentic MoVA / gating stack
32B denseLargest dense model in family
7.0B / 3.7BMid-size reasoning and scientific coding
0.9BUltra-small edge and tool-use focused
table · 2 cols
Transparency claimWhat IFM says ships
LLM360 standardCheckpoints per stage, data, logs, eval code
Training stacksxLLM (dense + MoE), RL360 post-training
Synthetic data~10T tokens with pipeline documentation
Shared scienceFour models on one 22T corpus for scaling studies

What changed versus older K2-Horizon mentions

explainx.ai first encountered K2-Horizon as secondhand social commentary around a rumored six-model drop. IFM’s November 2026 materials close that gap:

  • A public sizing ladder with deployment guidance per parameter class.
  • Architecture notes on MoVA and softplus attention gating in the paper abstract.
  • Uno integrated into the fleet narrative—extending the standalone Uno speedup release we documented in September.
  • Links to Hugging Face, vLLM, SGLang, and Ollama paths.

The IFM K2 timeline places Horizon after K2 Think V2 (January 2026) in the institute’s public lineage, signaling this is a generation refresh with heavier openness obligations, not a silent weight bump.

K2-Horizon open-weight release as seeds leaving a shared box toward builder handsK2-Horizon open-weight release as seeds leaving a shared box toward builder hands

Modeling ideas builders will actually touch

Mixture-of-Value Attention (MoVA)

Standard MoE routes feed-forward experts. IFM’s MoVA applies mixture-of-experts logic to attention value projections, aiming to strengthen reasoning without growing every active parameter. The paper claims a 36B-A4B MoVA model nearly matches a 32B dense model trained on the same data while activating about one-seventh the parameters—a claim to validate on your agent harness, not in a leaderboard screenshot alone.

Softplus attention gating

IFM describes softplus gating as a more flexible, theoretically grounded alternative to hard attention switches. For practitioners, the practical question is whether gating stabilizes long-horizon tool loops; IFM positions Horizon models for deep research and office automation alongside coding.

Uno decoding adapters

Uno adds a lossless diffusion adapter to speed decoding up to 2.2× without maintaining a separate speculative draft model. IFM already shipped K2-Horizon-7B-Uno weights on Hugging Face earlier; Horizon folds Uno into the fleet story rather than an isolated technique demo.

The openness package: what to clone first

IFM’s marketing uses "glass box, not black box" language. Concretely, the project page promises:

table · 2 cols
ArtifactWhy builders care
Pretraining corpusAudit data mix and contamination risk
Stage checkpointsStudy collapse, emergent capabilities, early stopping
xLLM codeReproduce dense and MoE training at production scale
RL360 stackCompare post-training recipes against open RL tooling you already run
Synthetic pipelinesReplicate or criticize 10T-token synthetic generation IFM highlights

That depth differentiates Horizon from weights-only drops that still dominate social feeds. It also raises hosting cost: publishing terabytes of data and checkpoints helps researchers but burdens mirrors—plan partial mirrors if you only need one size class.

K2-Horizon checkpoint branches showing staged training lineage for open MoE modelsK2-Horizon checkpoint branches showing staged training lineage for open MoE models

Performance claims and honest benchmarking

IFM states Horizon models lead or stay competitive with strong open models at each size class on agentic and coding benchmarks. The Hugging Face paper page does not replace a harness you trust—run Horizon checkpoints against your repo tests, agent evals, and cost caps.

Comparison anchors elsewhere on explainx.ai:

  • Marin Mochi 535B open MoE pretraining narrative for another full-stack openness pitch in 2026.
  • DeepSeek free-tier limits for how API economics push teams back toward self-hosted open weights.
  • Baseten frontier inference guide when you operationalize large MoE on rented GPUs.

How to try K2-Horizon this week

  1. Pick a size class — start 7B or 32B if you lack a 375B rack; use 0.9B only for edge prototypes IFM documents.
  2. Read the data card — IFM’s released corpus metadata matters for license and redistribution if you fine-tune commercially.
  3. Use listed runtimes — IFM names vLLM, SGLang, and Ollama; match your existing local inference workflows before porting agents.
  4. Log active parameters — MoE models punish naive VRAM planning; MoVA adds another knob versus dense baselines.
  5. Compare Uno on/off — if latency dominates, reproduce IFM’s Uno adapter path instead of bolting generic speculative decoding.

Limits and verification gaps

Enterprise flagship load. A 375B-A23B MoE is not a laptop download. Treat IFM’s headline as infrastructure news unless you operate a cluster or a host that already serves comparable open MoE.

Benchmark sourcing. Until independent labs publish harness logs, treat leaderboard cells as vendor-reported unless IFM linked raw eval artifacts you reproduced.

Geopolitical hosting. IFM is Abu Dhabi–based; some teams must map data residency and export rules before mirroring full corpora—separate from model quality.

Name collision. “K2-Horizon” appeared in rumors before primary docs existed. Anchor citations on ifm.ai/k2 and the 2611.21003 paper slug, not repost threads.

What developers are asking

table · 2 cols
QuestionPractical answer
Is this Apache or a custom license?Read the license files bundled with each checkpoint on Hugging Face before production use—we do not substitute for legal review here.
Does Horizon replace K2-Think?IFM’s timeline keeps Think as a prior line; Horizon is a new fleet with overlapping branding—pick weights by release tag, not name alone.
Synthetic 10T tokens—real or marketing?IFM says pipelines and documentation ship to verify; skepticism is healthy until you inspect recipes.

Related reading

  • IFM Uno lossless diffusion speedups on K2-Horizon-7B
  • Abu Dhabi IFM six-model rumor vs primary docs
  • Qwen Code open terminal agent trending on GitHub
  • Baseten efficient frontier LLM inference
  • Marin Mochi open MoE pretraining transparency
  • Loop engineering for coding agents
  • DeepSeek compute limits pushing self-hosting

Model-by-model: who each Horizon size is for

IFM’s public sizing page is unusually explicit about deployment intent. Summarizing the narrative (not a substitute for reading IFM’s own cards):

375B-A23B MoE — Positioned as the enterprise flagship for coding and agentic workloads where latency and dollar-per-task matter less than ceiling capability. Expect multi-GPU or hosted inference; this is not the model you quantize for a laptop experiment.

36B-A4B MoE — IFM’s efficiency play: MoVA and gating aim for agentic quality with far fewer active parameters per forward pass. Teams comparing cost per successful tool loop should benchmark this against dense 32B–70B rivals on identical prompts.

32B dense — The largest dense checkpoint in the ladder. Dense models remain easier to reason about for fine-tuning shops that do not want to debug expert routing on day one.

7.0B and 3.7B — Mid-tier options IFM ties to scientific coding and math at sizes that still fit serious fine-tuning jobs. Useful when you need Horizon’s data recipe transparency without Horizon’s hosting bill.

0.9B — IFM mentions watches, glasses, and constrained environments—a rarity among “frontier fleet” launches. Treat it as a probe for on-device tool use, not a replacement for phone-class models from larger vendors.

Synthetic data at 10T tokens: why builders should read the pipelines

Horizon’s abstract emphasizes pipeline documentation for synthetic pretraining—roughly 10T tokens generated through disclosed flows. That is a direct response to critics who say open releases hide data laundering behind opaque mixtures.

If you are building domain-specific corpora, IFM’s release is a template for what “open” might mean next: not only dedup stats, but generators, filters, and stage logs. The burden on you is time: auditing pipelines takes longer than pulling weights.

Pair that transparency debate with China–US distillation policy fights: synthetic and distilled data are now geopolitical objects, not ML trivia.

xLLM and RL360: infrastructure you can diff against your stack

Horizon bundles xLLM for dense and MoE training and RL360 for post-training. Even if you never rerun full pretraining, these repos matter when you ask:

  • Does IFM’s MoE routing match your internal trainer assumptions?
  • How does RL360 schedule tool-use rewards compared with loop engineering patterns you already run on Claude Code or Codex?

Open training code lets you diff recipes, not only download logits.

Risk checklist before production

table · 2 cols
RiskMitigation
License ambiguity on data shardsRead HF license + data card before commercial fine-tunes
MoE serving cost spikesCap concurrent experts; monitor active-parameter billing on hosts
Benchmark overfittingHold out your own agent eval suite; ignore leaderboard screenshots
Mirror bandwidthStage only the size class you need; avoid terabyte mirrors “just in case”
Security reviews on Abu Dhabi originMap procurement rules separately from model quality review

What people search after a Horizon announcement

table · 2 cols
Query shapeAnswer in one line
K2 Horizon vs K2 ThinkHorizon is the November 2026 fleet release with six sizes; Think lines are earlier generations on IFM’s timeline
Is K2 Horizon on Ollama?IFM lists Ollama among deployment paths on ifm.ai/k2—verify model tags before scripting pulls
LLM360 meaningIFM cites LLM360-style openness: checkpoints, data, logs, and code—not weights-only

Model cards, checkpoint counts, and inference partner listings change as IFM updates ifm.ai/k2. This article reflects materials linked from that site and Hugging Face paper 2611.21003 on November 17, 2026 UTC.

Spotted something out of date? Let us know.
Daily briefFrom Yash · 350K+ students taught

Catch up on AI

One AI story explained in plain English, plus one thing to try. Every day, free. Unsubscribe anytime.

  • ✓ Free
  • ✓ One email a day, no spam
  • ✓ Unsubscribe anytime
Yash Thakker

Written by

Yash Thakker

Yash is an AI expert with over 300K learners. Join his workshops →

View Yash Thakker in People in AI →

Related posts

Nov 17, 2026

Sage S2.1 Flash: Shanghai AI Lab Science Multimodal Agent

Shanghai AI Laboratory published Sage S2.1 Flash on November 16, 2026—a science multimodal agent with weights on Hugging Face and a public benchmark table spanning coding, lab tools, and search.

Nov 16, 2026

Echelon Optimizer: 10× Faster Muon Steps With a Lean Proof

On November 16, 2026 Evolving Machines released Echelon, an optimizer that approximates Muon polar updates with low-rank work. The post claims 10.5× faster optimizer steps on 70B-class layers, 14% shorter 1.3B training steps with matched validation loss, and large MoE savings on DeepSeek V4.1 Flash—with proofs checked in Lean.

Nov 13, 2026

Edge0-145B: Open 145B Agent Model Built to Run on Your Laptop

Edge0 released open Edge0-145B and Edge0-120B weights on November 12, 2026, targeting a single NVIDIA GPU with edge0-strata. The announcement claims frontier-class coding and agent scores at roughly 29GB peak memory on laptop-class hardware.