Skip to content

Open-weight models

Rising · emerged February 2023 · high August 2026 · +6% on the month before · 305 mentions on Trending in 14 days

Models whose trained weights anyone can download, run and fine-tune: Llama, Qwen, DeepSeek, Mistral and more. Named by Touvron et al., Meta AI: LLaMA: Open and Efficient Foundation Language Models (Feb 24, 2023). Also called Open weights, Open models, Open-source LLMs, Open-weights.

Index 0 to 100, monthly
04183202420252026Origin

Signals

SignalReads as
Hacker News stories, last 12 months379 stories, +206% on the 12 months beforeAdoption
arXiv papers, last 12 months1,593 papers, +300% on the 12 months beforeAdoption
Wikipedia views, August 20267,800, +126% on a year beforeAdoption
Product releases naming it11 since Sep 1, 2025Adoption
GitHub repositories1,458Adoption
Funding rounds naming it3, $1.7B in allAdoption

Trending terms

TermPeak
Qwen6 days, last Sep 24 yesterday30
Hugging Face8 days, last Sep 21 4 days ago23
Mistral AI2 days, last Sep 16 9 days ago15
DeepSeek9 days, last Sep 14 11 days ago33

Companies

Releases

11 releases since Sep 1, 2025
Name
Kimi K3 by Moonshot AI is now generally available on Amazon BedrockAWS · Sep 18, 2026 7 days ago
Announcing Cohere's North Small TranslateCohere · Sep 9, 2026 2 weeks ago
GLM 5.2 free for eve agents through August 27 via Blackbox on AI GatewayVercel · Aug 13, 2026 6 weeks ago
In-region inference, open models, and new European infrastructure for sovereign AIMistral AI · Aug 11, 2026 6 weeks ago
Agent Runners adds OpenCode and AI model controlsNetlify · Aug 6, 2026 7 weeks ago
Introducing ShieldstralMistral AI · Aug 4, 2026 7 weeks ago
RolloutBlack Forest Labs · Jul 23, 2026 2 months ago
Laguna S 2.1 is now available on AI GatewayVercel · Jul 21, 2026 2 months ago

Models

Papers

Name
The Llama 3 Herd of ModelsMeta · Jul 31, 2024 2 years ago
Nemotron-4 340B Technical ReportNVIDIA · Jun 17, 2024 2 years ago
LLaMA: Open and Efficient Foundation Language ModelsMeta · Feb 27, 2023 3 years ago

GitHub projects

1,458 repositories
NameStars
openai/gpt-ossgpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI · Jun 23, 2025 1 year ago20,427
oumi-ai/oumiEasily fine-tune, evaluate and deploy Qwen, Gemma, or any open weight LLM! · May 7, 2024 2 years ago9,390
Zyphra/ZonosZonos-v0.1 is a leading open-weight text-to-speech model trained on more than 200k hours of varied multilingual speech, delivering expressiv · Feb 7, 2025 1 year ago7,234
google-deepmind/gemmaGemma open-weight LLM library, from Google DeepMind · Feb 20, 2024 2 years ago5,746
MiniMax-AI/MiniMax-M1MiniMax-M1, the world's first open-weight, large-scale hybrid-attention reasoning model. · Jun 14, 2025 1 year ago3,184
AtomicBot-ai/atomic-agentAtomic Agent is a local-first AI agent. Runs open-weight models on your own machine via llama.cpp. · Apr 21, 2026 5 months ago2,494

Funding

3 rounds
NameAmount
Fireworks AISeries D · Jul 15, 2026 2 months ago$1,505M
Arcee AISeries B · Sep 16, 2026 9 days ago$150M
OllamaSeries B · Jul 9, 2026 2 months ago$65M
MovementStage
Local AIRunning language and image models on your own laptop, phone or server instead of a cloud API.Peak
Small language modelsCompact models, a few billion parameters or less, trained to be good enough and cheap to run anywhere.Peak
Mixture of expertsModels split into many expert blocks with only a few active per token, so they grow in size without growing in cost.Peak

History

DateWhat changed
Sep 24, 2026Tracking started: emerged February 2023, rising, 4 other names recorded

Sources: Curve: Hacker News story titles, Wikipedia pageviews and arXiv papers by month, refreshed Sep 24, 2026. Related entities from the fru.dev sites' public APIs, matched by name. How stages work.

Trending terms by email

Monday mornings: the week's top trending terms in data, tech and AI.

Double opt-in. Unsubscribe any time.