Skip to main content
Rase AI Intelligence Newsletter

Explore AI developments

Search published AI tools, models, robots, and research. Filter by organization, geography, milestones, and evidence.

Filters and sorting
Filter by milestone dates
Announcement
First access
Commercial release
Clear filters

66 profiles

Models & assistants

Hy4

Tencent (Hunyuan)

It is a language model family with documented preview and subsequent variants.

Medium confidenceEvidence checked

Models & assistants

Qwen4-Max

Alibaba (Qwen team, Alibaba Cloud)

It provides model-based assistance for text and related work tasks.

High confidenceEvidence checked

Models & assistants

Gemini 4 Hydrogen

Google DeepMind

It processes text and other supported inputs for tasks such as code assistance and planning.

High confidenceEvidence checked

Models & assistants

ERNIE 5.5 Flash Preview

Baidu

It is designed to process supported media inputs alongside text in the documented preview environment.

High confidenceEvidence checked

Models & assistants

Claude Fable 5.2 and Claude Mythos 5.2

Anthropic

The models assist with language and coding tasks under their documented access conditions.

High confidenceEvidence checked

Models & assistants

Muse Ember

Meta (AI at Meta)

It supports language and coding workflows through the interfaces described by Meta.

High confidenceEvidence checked

Models & assistants

K-EXAONE 2.5

LG AI Research

It provides a text reasoning model through a developer API.

High confidenceEvidence checked

Models & assistants

Jais 3

Inception42 and MBZUAI Institute of Foundation Models

It is designed to process Arabic and help with text-related tasks in the uses described by its developer.

Medium confidenceEvidence checked

Models & assistants

Sarvam 105B (open weights)

Sarvam AI

It is a language model released with downloadable weights and evaluated on Indian-language tasks.

High confidenceEvidence checked

Models & assistants

Latam-GPT

CENIA (Centro Nacional de Inteligencia Artificial)

Provides a language model intended to support work reflecting Latin American and Caribbean contexts.

High confidenceEvidence checked

Models & assistants

HUMAIN Chat

HUMAIN (PIF)

Supports conversational Arabic and English, with speech input across supported Arabic dialects and web search.

High confidenceEvidence checked

Models & assistants

A.X 4.0

SK Telecom

It generates responses to text instructions, with an emphasis on Korean-language tasks.

Medium confidenceEvidence checked

Models & assistants

Magistral

Mistral AI

The models generate responses to tasks involving step-by-step reasoning.

High confidenceEvidence checked

Models & assistants

Sarvam-M

Sarvam AI

A Sarvam AI language model designed for Indian-language tasks, with reasoning and non-reasoning generation modes described by the developer.

Medium confidenceEvidence checked

Models & assistants

Qwen3

Alibaba (Qwen team)

An Alibaba Qwen model family with dense and mixture-of-experts variants, including checkpoints that can switch between reasoning and ordinary responses.

Medium confidenceEvidence checked

Models & assistants

EXAONE 3.5

LG AI Research

The models generate responses to text instructions in the variants published by LG AI Research.

Medium confidenceEvidence checked

Models & assistants

Solar Pro

Upstage

It generates text responses for the documented developer use cases.

Medium confidenceEvidence checked

Models & assistants

Qwen2.5

Alibaba (Qwen team)

The family provides models for text generation, including versions tailored to different developer use cases.

High confidenceEvidence checked

Models & assistants

Jamba

AI21 Labs

It generates text using the hybrid architecture described by AI21.

High confidenceEvidence checked

Models & assistants

Le Chat (now Vibe)

Mistral AI

The assistant generates conversational responses from user prompts through Mistral's interface.

High confidenceEvidence checked

Models & assistants

Mistral 7B

Mistral AI

Generates text from prompts and can be deployed or fine-tuned using the published model weights.

High confidenceEvidence checked

Models & assistants

ChatGPT

OpenAI

It generates conversational text in response to prompts and follow-up messages.

High confidenceEvidence checked

Models & assistants

Ling-1T

Ant Group (inclusionAI)

A language model from inclusionAI that uses a mixture-of-experts architecture, with the developer positioning it for responses without a separate reasoning mode.

Medium confidenceEvidence checked

Models & assistants

N-ATLaS

National Centre for Artificial Intelligence and Robotics (NCAIR, Nigeria)

A Nigeria-focused language model effort with a Llama-based text checkpoint supporting Hausa, Igbo, Yoruba and Nigerian English.

Medium confidenceEvidence checked