目次
テキストモデル
| モデル名 | 最大トークン数 | 言語 | ユースケース | 特徴 |
|---|---|---|---|---|
|
Claude 2 Anthropic |
100K | English and multiple other languages | Thoughtful dialogue, content creation, complex reasoning, creativity, and coding. Claude 2 excels at the core capabilities of Claude 1.3 and Claude Instant. | Claude 2 is a general purpose large language model, and the most capable system released by Anthropic to date. |
|
Claude 1.3 Anthropic |
100K | English and multiple other languages | Search, writing, editing, outlining, and summarizing text; coding; and providing helpful advice about a broad range of subjects. | Claude 1.3 is an earlier version of Anthropic's general purpose large language models. |
|
Claude Instant Anthropic |
100K | English and multiple other languages | Casual dialogue, text analysis, summarization, and document comprehension. | Claude Instant is Anthropic's faster, lower-priced yet very capable large language model. |
|
Command Cohere |
4k | English | Chat, text generation, text summarization. | Command is Cohere’s generative large language model (52B parameters). |
|
Jurassic-2 Ultra AI21 Labs |
8,192 | English, Spanish, French, German, Portuguese, Italian, and Dutch | Question answering, summarization, draft generation, advanced information extraction, ideation for tasks requiring intricate reasoning and logic. | AI21’s most powerful model, offering exceptional quality. |
|
Jurassic-2 Mid AI21 Labs |
8,192 | English, Spanish, French, German, Portuguese, Italian, and Dutch | Question answering, summarization, draft generation, advanced information extraction, ideation. | AI21’s mid-sized model, designed to strike the right balance between exceptional quality and affordability. |
|
Titan Text G1 - Express AWS |
8k | English | Open ended text generation, brainstorming, summarization, code generation, table creation, data formatting, paraphrasing, chain of though, rewrite, extraction, Q&A, chat | Express is a generative large language model (LLM) for tasks such as summarization, text generation (for example, creating a blog post), classification, open-ended Q&A, and information extraction. |
|
Llama-2-13b-chat Meta |
4K | English | Assistant-like chat | Fine-tuned model in the parameter size of 13B. |
|
Llama-2-70b-chat Meta |
4K | English | Assistant-like chat | Fine-tuned model in the parameter size of 70B. |
日本語で使うならAnthropicのClaudeが良さそう。100Kトークンもすごい。
Llama-2-13b-chatとLlama-2-70b-chatは2023/10/1時点で未提供
テキストモデル（Embeddings）
| モデル名 | 最大トークン数 | 言語 | ユースケース | 特徴 |
|---|---|---|---|---|
|
Titan Embeddings Generation 1 (G1) AWS |
8k | Multilingual | text retrieval, semantic similarity, clustering | Text is fast and cost-effective. Designed for fast responses, this model is an ideal choice for language processing tasks that require maximum affordability and less processing power. |
Titan Embeddingsは日本語にも対応。
CohereはMultilingual Embed Modelがあるので、そのうち提供されるかも
画像モデル
| モデル名 | 最大トークン数 | 言語 | ユースケース | 特徴 |
|---|---|---|---|---|
|
Stable Diffusion XL 1.0 Stability AI |
77 token limit for prompts | English | Advertising & Marketing, Media & Entertainment, Gaming & Metaverse. | The most advanced text to image model from Stability AI. |
|
Stable Diffusion XL 0.8 Stability AI |
77 token limit for prompts | English | Advertising & Marketing, Media & Entertainment, Gaming & Metaverse. | Text to image model from Stability AI. |
Stable Diffusion XL 1.0は2023/10/1時点で未提供
参考
https://aws.amazon.com/jp/bedrock/claude/
https://aws.amazon.com/jp/bedrock/cohere-command/
https://aws.amazon.com/jp/bedrock/jurassic/
https://aws.amazon.com/jp/bedrock/llama-2/
https://aws.amazon.com/jp/bedrock/stable-diffusion/
https://aws.amazon.com/jp/bedrock/titan/