Bright

Find your way in.

106 items for “open models”: records, experiences, guides and model families. Source reports remain distinct from discoveries.

guide · Source published date unknown

Open Models — intelligence you can build on

A source-checked guide to open weights, openness, model families, licensing, local hardware, applications and open-versus-hosted tradeoffs.

guide · Source published date unknown

Explore open-model families by intended use

Bright’s atlas is a curated set of checked model-family records. Its entry paths identify relevant families to explore for local use, coding, multimodal work, robotics, science, research or licensing; they are not rankings or “best model” recommendations.

discovery · Source published 2026-03-11

An open reasoning model with its recipe beside it.

NVIDIA released Nemotron 3 Super, an open 120-billion-parameter mixture-of-experts reasoning model with 12 billion active parameters, plus stated releases of its methodology, data, reinforcement-learning environments, and evaluation recipes.

discovery · Source published 2026-09-09

When a model changed, its open recipe helped trace why.

Goodfire used Ai2's open OLMo post-training stack to trace a known regression and inspect behavioral shifts.

discovery · Source published 2026-09-10

An open model learns the Moon's surface.

NASA and IBM released a Lunar Foundation Model trained primarily on Lunar Reconnaissance Orbiter data for research tasks such as crater and volcanic-feature mapping and possible polar-ice analysis.

model-family · Source published date unknown

gpt-oss: license, size and what is open

Two open-weight text reasoning models with adjustable reasoning effort, structured output, and tool-use support.

model-family · Source published date unknown

Olmo 3: license, size and what is open

A fully open model flow spanning base, mid-trained, long-context, instruction, reasoning, and RL-zero checkpoints.

model-family · Source published date unknown

Molmo 2: license, size and what is open

A family of open vision-language models for images, multiple images, and video, with explicit spatial outputs such as pointing and object tracking.

model-family · Source published date unknown

Isaac GR00T: license, size and what is open

An open vision-language-action model for generalized humanoid skills, accompanied by reference code, data tooling, and a robotics development stack.

discovery · Source published 2026-10-03

Kolibri gives teams another open-weight AI option. Here’s what it takes to run it.

Germany’s Aleph Alpha released a downloadable German-English model on October 3. It offers another route to self-hosted AI, while showing why open weights, open-source development and easy local use are different things.

guide · Source published date unknown

What “open source” means for space-science AI

A model can be downloadable without making every ingredient open. Here is how to inspect the code, weights, data, license and evidence separately.

discovery · Source published 2026-10-15

CERN and Mistral plan a physics model for the next era of collider data

A statement of intent opens a new research effort. Existing particle-reconstruction work shows why the idea is promising, and how much evidence a broader model will need.

collection · Authored Bright experience

Open-Source & Open-Weight AI Models: Compare and Run Locally

Compare 15 AI model families by license, size and openness. Inspect documented local candidates, try a sourced setup guide, and download a dated comparison.

model-family · Source published date unknown

DeepSeek V4.1 Flash: license, size and what is open

A downloadable DeepSeek multimodal Flash model, with sparse activation that differs between prompt processing and generation.

model-family · Source published date unknown

Mistral Small 4: license, size and what is open

A hybrid sparse model that combines instruction following, reasoning, visual understanding, and coding modes in one downloadable checkpoint.

model-family · Source published 2026-04-27

Nemotron 3 Nano Omni: license, size and what is open

A compact sparse reasoning model that accepts video, audio, images, and text while activating about three billion parameters per token.

model-family · Source published date unknown

SmolLM3: license, size and what is open

A compact multilingual reasoning model whose engineering blueprint, mixtures, training frameworks, and alignment process are published.

model-family · Source published date unknown

Kimi K2.5: license, size and what is open

A trillion-parameter visual agentic model with a comparatively small active footprint and public weights and code.

model-family · Source published date unknown

Evo 2: license, size and what is open

A biological foundation model trained across genomes for sequence understanding and design, released with code, weights, and OpenGenome2 data.

model-family · Source published date unknown

Voxtral Realtime: license, size and what is open

A compact streaming speech-recognition model designed for low-latency transcription through a realtime API.