MODEL DIRECTORY

Latest AI models

New releases, in one place. Explore models by their publication date.

Newest first7 languages
Clear

641 models

Newest first
DeepSeek

DeepSeek Coder

2023-11-02 · DeepSeek A family of several model sizes for code completion, generation and programming questions. The family name is not a unique API snapshot. The model name is used instead of a…

Alibaba

通义千问 2.0(商业版)

2023-10-31 · Alibaba A commercial language-model update announced at Apsara Conference for questions, writing and enterprise applications. It was not the next year's open-weight Qwen2. The model name…

Zhipu AI

ChatGLM3-6B

2023-10-27 · Zhipu AI The open-weight dialogue variant of third-generation ChatGLM. Its new prompt format supported multi-turn conversation and structured tool calls. The model name is used instead of…

Baidu

文心大模型 4.0

2023-10-17 · Baidu Introduced at Baidu World with invited testing, this version targeted understanding, generation, reasoning and multi-turn interaction. This was not ERNIE Bot's first release. The…

Moonshot AI

Kimi Chat

2023-10-09 · Moonshot AI A Chinese assistant for long-document reading that supported about 200,000 Chinese characters at launch. The date refers to the assistant product, not the later API. The model…

Tencent

腾讯混元

2023-09-07 · Tencent Tencent introduced its general-purpose model at its Global Digital Ecosystem Summit. Companies in China could test and integrate it through Tencent Cloud's API. This was separate…

Anthropic

Claude Instant 1.2

2023-08-09 · Anthropic This API update to the fast-response Instant series improved information extraction, coding and adherence to requested output formats. Source

Anthropic

Claude 2

2023-07-11 · Anthropic The second generation of Claude supported long-document reading, coding and analysis. API access accompanied a public web beta in the US and UK. The model name stands in for an…

Zhipu AI

ChatGLM2-6B

2023-06-25 · Zhipu AI The second-generation bilingual dialogue model retained local deployment while improving conversation and support for longer inputs. The model name is used instead of a verified…

MiniMax

abab5.5

2023-05-31 · MiniMax A MiniMax text model in the abab family. The prospectus dates its release to May 2023 without giving a day; month-end is a placeholder. The model name is used instead of a…

Google

PaLM 2

2023-05-10 · Google Google announced this successor to PaLM at I/O for multilingual tasks, reasoning and coding. The date marks the family announcement, not a particular endpoint launch. The model…

Baidu

文心一言(ERNIE Bot)

2023-03-16 · Baidu Baidu demonstrated a Chinese-language assistant based on ERNIE for questions and content creation. Invited testing began that day; public access in mainland China followed on August…

Zhipu AI

ChatGLM-6B

2023-03-14 · Zhipu AI A bilingual Chinese-English dialogue model with weights available for local deployment. The project already provided a multi-turn example and model downloads on this date. The…

OpenAI

GPT-4

2023-03-14 · OpenAI A model for complex questions, writing and coding. Access through ChatGPT Plus began at launch; later expansion of API access was not a separate model. Source 1 · Source 2

Anthropic

Claude Instant

2023-03-14 · Anthropic Announced alongside the original Claude, Instant was designed for conversational applications that needed faster responses. The model name stands in for an unverified API…

Anthropic

Claude

2023-03-14 · Anthropic The original Claude was announced for tasks including question answering, summarization and writing. Claude Instant, announced the same day, was a separate version. The model…

OpenAI

ChatGPT(GPT-3.5 系列研究预览)

2022-11-30 · OpenAI The GPT-3.5-based web chat product opened as a research preview, supporting follow-up questions and revisions. This was not the GPT-3.5 Turbo API launch. The model name stands in…

Zhipu AI

GLM-130B

2022-08-04 · Zhipu AI A 130-billion-parameter pretrained foundation model for Chinese and English. It was not ChatGLM-6B, the dialogue model released the following year. The model name is used instead…

Baidu

文心 ERNIE 3.0 Zeus

2022-05-20 · Baidu Baidu presented this hundred-billion-parameter pretrained model at WAVE SUMMIT. It integrated task-related knowledge into language modeling and was distinct from the later ERNIE Bot…

MiniMax

abab1

2022-04-30 · MiniMax MiniMax’s first text model served internal products. The prospectus confirms April 2022 only; the last day of the month is a placeholder, not a public API launch date. The model…