Research
Everything Intel keeps, in one place — every brief we run, every department, readers' own, and every finding behind them. Browse below, or see the same body of work as a connected graph on the live map.
Following opens with early access — sign up for early access to register your interest; it does not switch following on today. Everything you follow will land on one page, in your account menu.
All coverage
Everything Intel has read, newest first. Each title opens at its original publisher.
- Open Source & Self-Hostable✓ verified
DeepSeek publishes V4 Pro 0813 weights at 1.7 trillion parameters
A 1.7T open-weights checkpoint is not something most teams will self-host, but it sets the ceiling that hosted providers price against — and the missing license line is the thing to check before it reaches anyone's build.
- Open Source & Self-Hostable
NVIDIA ships a 30B Nemotron for agent sub-tasks plus an open router that cuts cost to roughly a third of Opus 4.8
The stated tradeoff is published rather than hidden — 21–58% cheaper for 6–28% less accurate is a decision a team can actually take per sub-task instead of per product.
- Open Source & Self-Hostable✓ verified
Meta releases Muse Glimmer, a 30B Apache-2.0 multimodal model built for local agentic use
A 30B Apache-2.0 model that runs agentic work on your own hardware puts a real floor under what you have to pay a frontier API for — and llama.cpp already ships tool-call handling for it.
- Open Source & Self-Hostable
MiniMax releases MiniMax-H3, an open-weights 33B image-to-video model — community tooling lands within a day
Open-weights video generation at this scale moves a capability that was API-only last year onto self-hosted rigs — and the day-one tooling shows real demand.
- Open Source & Self-Hostable✓ verified
NVIDIA opens Alpamayo 2 Super, a 34B vision-language-action model for robotaxis, to commercial use
An openly licensed driving-grade vision-language-action model lowers the entry cost for autonomy builders — and shows NVIDIA competing on models, not just chips.
- Open Source & Self-Hostable
Mistral releases Shieldstral, a 3B open-weights policy-adaptive safety classifier
Guardrails you can rewrite in plain language without retraining — and self-host on commodity hardware — change the cost calculus for anyone shipping user-facing AI.
- Open Source & Self-Hostable✓ verified
Liquid AI releases LFM2.5-2.6B, an open-weights on-device model built for local agents
Tool-calling agents that run entirely on a device change the privacy and cost math for shipping AI features — no per-token bill and no data leaving the machine.
- Open Source & Self-Hostable
Cursor open-sources Mixture-of-Kittens, a deterministic MoE training megakernel for NVL72 racks
An application company open-sourcing frontier-grade training kernels signals how far serious model infrastructure has spread beyond the big labs.
- Open Source & Self-Hostable✓ verified
llama.cpp merges multi-token prediction for DeepSeek V4 — local decode roughly 50% faster
Self-hosted inference of the newest hot open-weights family just got materially faster on consumer hardware — worth re-benchmarking your local stack.
- Open Source & Self-Hostable✓ verified
LG ships K-EXAONE 2.0 — Korea's largest open model, 750B parameters under Apache 2.0
Sovereign-AI programs are now shipping frontier-scale open weights under permissive licenses — the self-hostable ceiling keeps rising.
- Open Source & Self-Hostable✓ verified
DeepSeek re-post-trains V4 Flash — and says the cheap model now beats its own flagship on agent work
Post-training alone moved a 13B-active open model past a flagship on agent benchmarks — the cheapest tier is where agent workloads may now land.
- Open Source & Self-Hostable
A 26-billion-parameter model running in about 2 GB of RAM on an 8 GB MacBook
The cheapest hedge against memory prices is needing less memory - this is the same squeeze that is moving chip stocks, showing up as an engineering constraint on the laptop you already own.
- Open Source & Self-Hostable
NVIDIA open-sources a surgical-robot simulator that runs 16x faster than the model it was distilled from
Surgical-robot policy work has been gated on scarce hardware time; a faster-than-real-time simulator with published weights moves that loop onto one workstation GPU.
- Open Source & Self-Hostable✓ verified
Moonshot puts the Kimi K3 weights on Hugging Face — under a licence that bills large model-as-a-service resellers
Open weights at frontier scale are only useful if you read the licence first: the revenue threshold is what decides whether self-hosting K3 is a free build or a contract negotiation.
- Open Source & Self-Hostable✓ verified
Debian puts LLM-assisted contributions to a project-wide vote — four proposals from outright ban to accept-with-responsibility
Whatever Debian settles on becomes the reference text other maintainer-run projects copy — and every one of the four options requires contributors to disclose, which turns provenance from an ethics debate into a workflow requirement.
- Open Source & Self-Hostable✓ verified
Upstage releases Solar Open 2 — a 250B open-weight model built for long-horizon agent work
A self-hostable, independently benchmarked alternative to closed agent APIs — runnable on two H200s when quantized.
- Open Source & Self-Hostable✓ verified
Poolside releases Laguna S 2.1 — an open-weight coding model that punches 10x above its size
A single-GPU-runnable open model competitive on real coding benchmarks lowers the hardware and cost bar for self-hosting a capable coding agent instead of paying frontier-API prices — and there's a free tier to try it today.
- Open Source & Self-Hostable✓ verified
Motif-3-Beta: a Korean startup ships a from-scratch 314B open model, not another derivative
Evidence that competitive from-scratch frontier-scale MoE design is diffusing beyond the established labs — worth tracking even though the license blocks commercial use for now.
- Open Source & Self-Hostable
Microsoft releases Mage-Flow — a compact, MIT-licensed image model that generates and edits at any resolution
A small, permissively licensed, single-GPU image model is a practical self-hosted alternative to closed image APIs for product teams.
- Open Source & Self-Hostable✓ verified
Meituan’s LongCat 2.0 — a 1.6-trillion-parameter open coder trained on Chinese chips — lands on OpenRouter
A near-frontier, MIT-licensed agentic coder at roughly a tenth of frontier pricing — and a supply-chain signal that trillion-parameter training no longer requires NVIDIA hardware.
- Open Source & Self-Hostable
LM Studio launches Bionic — an AI agent built for open models, with fully local voice input
Agentic workflows just reached the local-first crowd — you can point an agent at real coding and document work without sending your data to a frontier lab.
- Open Source & Self-Hostable✓ verified
Thinking Machines releases Inkling — its first open-weights model, a near-trillion-parameter multimodal base built for fine-tuning
A frontier-lab open base model designed for fine-tuning changes the build-vs-rent calculus for any team that needs a domain-specialized model.
- Open Source & Self-Hostable
Ollama 0.32 turns the CLI into an agent — with a cloud model as the default
If your self-hosted stack shells out to `ollama`, the default behavior just changed — and the deprecation list says which local models to migrate off.
- Open Source & Self-Hostable✓ verified
Bonsai 27B: the first 27B-class open model that runs on a phone
A 27B-class model in 3.9 GB moves serious multi-step agent work onto hardware you already own — no API bill, no data leaving the device.