@promptbook/openai 0.94.0-7 → 0.95.0

package/README.md

@@ -25,10 +25,6 @@ Write AI applications using plain human language across multiple models and plat
 
 
 
-<blockquote style="color: #ff8811">
-    <b>⚠ Warning:</b> This is a pre-release version of the library. It is not yet ready for production use. Please look at <a href="https://www.npmjs.com/package/@promptbook/core?activeTab=versions">latest stable release</a>.
-</blockquote>
-
 ## 📦 Package `@promptbook/openai`
 
 - Promptbooks are [divided into several](#-packages) packages, all are published from [single monorepo](https://github.com/webgptorg/promptbook).
@@ -112,11 +108,11 @@ console.info(outputParameters);
 Run books without any settings, boilerplate or struggle in Node.js:
 
 ```typescript
-import { wizzard } from '@promptbook/wizzard';
+import { wizard } from '@promptbook/wizard';
 
 const {
     outputParameters: { joke },
-} = await wizzard.execute(`https://github.com/webgptorg/book/blob/main/books/templates/generic.book`, {
+} = await wizard.execute(`https://github.com/webgptorg/book/blob/main/books/templates/generic.book`, {
     topic: 'Prague',
 });
 
@@ -259,17 +255,15 @@ Rest of the documentation is common for **entire promptbook ecosystem**:
 
 ## 🤍 The Book Abstract
 
-**It's time for a paradigm shift! The future of software is in plain English, French or Latin.**
+**It's time for a paradigm shift! The future of software is written in plain English, French, or Latin.**
 
 During the computer revolution, we have seen [multiple generations of computer languages](https://github.com/webgptorg/promptbook/discussions/180), from the physical rewiring of the vacuum tubes through low-level machine code to the high-level languages like Python or JavaScript. And now, we're on the edge of the **next revolution**!
 
-
-
 It's a revolution of writing software in **plain human language** that is understandable and executable by both humans and machines – and it's going to change everything!
 
 The incredible growth in power of microprocessors and the Moore's Law have been the driving force behind the ever-more powerful languages, and it's been an amazing journey! Similarly, the large language models (like GPT or Claude) are the next big thing in language technology, and they're set to transform the way we interact with computers.
 
-This shift is going to happen, whether we are ready for it or not. Our mission is to make it excellently, not just good.
+This shift will happen whether we're ready or not. Our mission is to make it excellent, not just good.
 
 **Join us in this journey!**
 
@@ -390,9 +384,6 @@ Join our growing community of developers and users:
 
 _A concise, Markdown-based DSL for crafting AI workflows and automations._
 
-
-
-
 ### Introduction
 
 Book is a Markdown-based language that simplifies the creation of AI applications, workflows, and automations. With human-readable commands, you can define inputs, outputs, personas, knowledge sources, and actions—without needing model-specific details.
@@ -420,7 +411,7 @@ Book is a Markdown-based language that simplifies the creation of AI application
 → {article}
 ```
 
-Each part of the book defines one of 3 circles:
+Each part of the book defines one of three circles:
 
 ### **1. What:** Workflows, Tasks and Parameters
 
@@ -442,8 +433,6 @@ Personas can have access to different knowledge, tools and actions. They can als
 
 -   [PERSONA](https://github.com/webgptorg/promptbook/blob/main/documents/commands/PERSONA.md)
 
-
-
 ### **3. How:** Knowledge, Instruments and Actions
 
 The resources used by the personas are used to do the work.
@@ -458,9 +447,9 @@ The resources used by the personas are used to do the work.
 
 Book language is based on markdown. It is subset of markdown. It is designed to be easy to read and write. It is designed to be understandable by both humans and machines and without specific knowledge of the language.
 
-The file has `.book` extension. It uses `UTF-8` non BOM encoding.
+The file has a `.book` extension and uses UTF-8 encoding without BOM.
 
-Book has two variants: flat - which is just a prompt with no structure, and full - which has a structure with tasks, commands and prompts.
+Books have two variants: flat — just a prompt without structure, and full — with tasks, commands, and prompts.
 
 As it is source code, it can leverage all the features of version control systems like git and does not suffer from the problems of binary formats, proprietary formats, or no-code solutions.
 
@@ -491,13 +480,13 @@ Or you can install them separately:
 
 -   ⭐ **[ptbk](https://www.npmjs.com/package/ptbk)** - Bundle of all packages, when you want to install everything and you don't care about the size
 -   **[promptbook](https://www.npmjs.com/package/promptbook)** - Same as `ptbk`
--   ⭐🧙‍♂️ **[@promptbook/wizzard](https://www.npmjs.com/package/@promptbook/wizzard)** - Wizzard to just run the books in node without any struggle
+-   ⭐🧙‍♂️ **[@promptbook/wizard](https://www.npmjs.com/package/@promptbook/wizard)** - Wizard to just run the books in node without any struggle
 -   **[@promptbook/core](https://www.npmjs.com/package/@promptbook/core)** - Core of the library, it contains the main logic for promptbooks
 -   **[@promptbook/node](https://www.npmjs.com/package/@promptbook/node)** - Core of the library for Node.js environment
 -   **[@promptbook/browser](https://www.npmjs.com/package/@promptbook/browser)** - Core of the library for browser environment
 -   ⭐ **[@promptbook/utils](https://www.npmjs.com/package/@promptbook/utils)** - Utility functions used in the library but also useful for individual use in preprocessing and postprocessing LLM inputs and outputs
 -   **[@promptbook/markdown-utils](https://www.npmjs.com/package/@promptbook/markdown-utils)** - Utility functions used for processing markdown
--   _(Not finished)_ **[@promptbook/wizzard](https://www.npmjs.com/package/@promptbook/wizzard)** - Wizard for creating+running promptbooks in single line
+-   _(Not finished)_ **[@promptbook/wizard](https://www.npmjs.com/package/@promptbook/wizard)** - Wizard for creating+running promptbooks in single line
 -   **[@promptbook/javascript](https://www.npmjs.com/package/@promptbook/javascript)** - Execution tools for javascript inside promptbooks
 -   **[@promptbook/openai](https://www.npmjs.com/package/@promptbook/openai)** - Execution tools for OpenAI API, wrapper around OpenAI SDK
 -   **[@promptbook/anthropic-claude](https://www.npmjs.com/package/@promptbook/anthropic-claude)** - Execution tools for Anthropic Claude API, wrapper around Anthropic Claude SDK 
@@ -541,9 +530,7 @@ The following glossary is used to clarify certain concepts:
 -   **Retrieval-augmented generation** is a machine learning paradigm where a model generates text by retrieving relevant information from a large database of text. This approach combines the benefits of generative models and retrieval models.
 -   **Longtail** refers to non-common or rare events, items, or entities that are not well-represented in the training data of machine learning models. Longtail items are often challenging for models to predict accurately.
 
-_Note: This section is not complete dictionary, more list of general AI / LLM terms that has connection with Promptbook_
-
-
+_Note: This section is not a complete dictionary, more list of general AI / LLM terms that has connection with Promptbook_
 
 ### 💯 Core concepts
 

package/esm/index.es.js

@@ -18,7 +18,7 @@ const BOOK_LANGUAGE_VERSION = '1.0.0';
  * @generated
  * @see https://github.com/webgptorg/promptbook
  */
-const PROMPTBOOK_ENGINE_VERSION = '0.94.0-7';
+const PROMPTBOOK_ENGINE_VERSION = '0.95.0';
 /**
  * TODO: string_promptbook_version should be constrained to the all versions of Promptbook engine
  * Note: [💞] Ignore a discrepancy between file name and entity name
@@ -514,7 +514,7 @@ function checkSerializableAsJson(options) {
         else {
             for (const [subName, subValue] of Object.entries(value)) {
                 if (subValue === undefined) {
-                    // Note: undefined in object is serializable - it is just omited
+                    // Note: undefined in object is serializable - it is just omitted
                     continue;
                 }
                 checkSerializableAsJson({ name: `${name}.${subName}`, value: subValue, message });
@@ -1295,7 +1295,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'COMPLETION',
             modelTitle: 'davinci-002',
             modelName: 'davinci-002',
-            modelDescription: 'Legacy completion model with strong performance on text generation tasks. Optimized for complex instructions and longer outputs.',
+            modelDescription: 'Legacy completion model with 4K token context window. Excels at complex text generation, creative writing, and detailed content creation with strong contextual understanding. Optimized for instructions requiring nuanced outputs and extended reasoning. Suitable for applications needing high-quality text generation without conversation management.',
             pricing: {
                 prompt: pricing(`$2.00 / 1M tokens`),
                 output: pricing(`$2.00 / 1M tokens`),
@@ -1313,7 +1313,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-3.5-turbo-16k',
             modelName: 'gpt-3.5-turbo-16k',
-            modelDescription: 'GPT-3.5 Turbo with extended 16k token context length for handling longer conversations and documents.',
+            modelDescription: 'Extended context GPT-3.5 Turbo with 16K token window. Maintains core capabilities of standard 3.5 Turbo while supporting longer conversations and documents. Features good balance of performance and cost for applications requiring more context than standard 4K models. Effective for document analysis, extended conversations, and multi-step reasoning tasks.',
             pricing: {
                 prompt: pricing(`$3.00 / 1M tokens`),
                 output: pricing(`$4.00 / 1M tokens`),
@@ -1337,7 +1337,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-4',
             modelName: 'gpt-4',
-            modelDescription: 'GPT-4 is a powerful language model with enhanced reasoning, instruction-following capabilities, and 8K context window. Optimized for complex tasks requiring deep understanding.',
+            modelDescription: 'Powerful language model with 8K context window featuring sophisticated reasoning, instruction-following, and knowledge capabilities. Demonstrates strong performance on complex tasks requiring deep understanding and multi-step reasoning. Excels at code generation, logical analysis, and nuanced content creation. Suitable for advanced applications requiring high-quality outputs.',
             pricing: {
                 prompt: pricing(`$30.00 / 1M tokens`),
                 output: pricing(`$60.00 / 1M tokens`),
@@ -1349,7 +1349,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-4-32k',
             modelName: 'gpt-4-32k',
-            modelDescription: 'Extended context version of GPT-4 with a 32K token window for processing very long inputs and generating comprehensive responses for complex tasks.',
+            modelDescription: 'Extended context version of GPT-4 with 32K token window. Maintains all capabilities of standard GPT-4 while supporting analysis of very lengthy documents, code bases, and conversations. Features enhanced ability to maintain context over long interactions and process detailed information from large inputs. Ideal for document analysis, legal review, and complex problem-solving.',
             pricing: {
                 prompt: pricing(`$60.00 / 1M tokens`),
                 output: pricing(`$120.00 / 1M tokens`),
@@ -1372,7 +1372,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-4-turbo-2024-04-09',
             modelName: 'gpt-4-turbo-2024-04-09',
-            modelDescription: 'Latest stable GPT-4 Turbo model from April 2024 with enhanced reasoning and context handling capabilities. Offers 128K context window and improved performance.',
+            modelDescription: 'Latest stable GPT-4 Turbo from April 2024 with 128K context window. Features enhanced reasoning chains, improved factual accuracy with 40% reduction in hallucinations, and better instruction following compared to earlier versions. Includes advanced function calling capabilities and knowledge up to April 2024. Provides optimal performance for enterprise applications requiring reliability.',
             pricing: {
                 prompt: pricing(`$10.00 / 1M tokens`),
                 output: pricing(`$30.00 / 1M tokens`),
@@ -1384,7 +1384,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-3.5-turbo-1106',
             modelName: 'gpt-3.5-turbo-1106',
-            modelDescription: 'November 2023 version of GPT-3.5 Turbo with improved instruction following and a 16K token context window.',
+            modelDescription: 'November 2023 version of GPT-3.5 Turbo with 16K token context window. Features improved instruction following, more consistent output formatting, and enhanced function calling capabilities. Includes knowledge cutoff from April 2023. Suitable for applications requiring good performance at lower cost than GPT-4 models.',
             pricing: {
                 prompt: pricing(`$1.00 / 1M tokens`),
                 output: pricing(`$2.00 / 1M tokens`),
@@ -1396,7 +1396,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-4-turbo',
             modelName: 'gpt-4-turbo',
-            modelDescription: 'More capable model than GPT-4 with improved instruction following, function calling and a 128K token context window for handling very large documents.',
+            modelDescription: 'More capable and cost-efficient version of GPT-4 with 128K token context window. Features improved instruction following, advanced function calling capabilities, and better performance on coding tasks. Maintains superior reasoning and knowledge while offering substantial cost reduction compared to base GPT-4. Ideal for complex applications requiring extensive context processing.',
             pricing: {
                 prompt: pricing(`$10.00 / 1M tokens`),
                 output: pricing(`$30.00 / 1M tokens`),
@@ -1408,10 +1408,10 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'COMPLETION',
             modelTitle: 'gpt-3.5-turbo-instruct-0914',
             modelName: 'gpt-3.5-turbo-instruct-0914',
-            modelDescription: 'September 2023 version of GPT-3.5 Turbo optimized for completion-style instruction following with a 4K context window.',
+            modelDescription: 'September 2023 version of GPT-3.5 Turbo Instruct with 4K context window. Optimized for completion-style instruction following with deterministic responses. Better suited than chat models for applications requiring specific formatted outputs without conversation management. Knowledge cutoff from September 2021.',
             pricing: {
                 prompt: pricing(`$1.50  / 1M tokens`),
-                output: pricing(`$2.00 / 1M tokens`), // <- For gpt-3.5-turbo-instruct
+                output: pricing(`$2.00 / 1M tokens`),
             },
         },
         /**/
@@ -1420,7 +1420,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'COMPLETION',
             modelTitle: 'gpt-3.5-turbo-instruct',
             modelName: 'gpt-3.5-turbo-instruct',
-            modelDescription: 'Optimized version of GPT-3.5 for completion-style API with good instruction following and a 4K token context window.',
+            modelDescription: 'Optimized version of GPT-3.5 for completion-style API with 4K token context window. Features strong instruction following with single-turn design rather than multi-turn conversation. Provides more consistent, deterministic outputs compared to chat models. Well-suited for templated content generation and structured text transformation tasks.',
             pricing: {
                 prompt: pricing(`$1.50  / 1M tokens`),
                 output: pricing(`$2.00 / 1M tokens`),
@@ -1438,7 +1438,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-3.5-turbo',
             modelName: 'gpt-3.5-turbo',
-            modelDescription: 'Latest version of GPT-3.5 Turbo with improved performance and instruction following capabilities. Default 4K context window with options for 16K.',
+            modelDescription: 'Latest version of GPT-3.5 Turbo with 4K token default context window (16K available). Features continually improved performance with enhanced instruction following and reduced hallucinations. Offers excellent balance between capability and cost efficiency. Suitable for most general-purpose applications requiring good AI capabilities at reasonable cost.',
             pricing: {
                 prompt: pricing(`$0.50 / 1M tokens`),
                 output: pricing(`$1.50 / 1M tokens`),
@@ -1450,7 +1450,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-3.5-turbo-0301',
             modelName: 'gpt-3.5-turbo-0301',
-            modelDescription: 'March 2023 version of GPT-3.5 Turbo with a 4K token context window. Legacy model maintained for backward compatibility.',
+            modelDescription: 'March 2023 version of GPT-3.5 Turbo with 4K token context window. Legacy model maintained for backward compatibility with specific application behaviors. Features solid conversational abilities and basic instruction following. Knowledge cutoff from September 2021. Suitable for applications explicitly designed for this version.',
             pricing: {
                 prompt: pricing(`$1.50 / 1M tokens`),
                 output: pricing(`$2.00 / 1M tokens`),
@@ -1462,7 +1462,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'COMPLETION',
             modelTitle: 'babbage-002',
             modelName: 'babbage-002',
-            modelDescription: 'Efficient legacy completion model with a good balance of performance and speed. Suitable for straightforward text generation tasks.',
+            modelDescription: 'Efficient legacy completion model with 4K context window balancing performance and speed. Features moderate reasoning capabilities with focus on straightforward text generation tasks. Significantly more efficient than davinci models while maintaining adequate quality for many applications. Suitable for high-volume, cost-sensitive text generation needs.',
             pricing: {
                 prompt: pricing(`$0.40 / 1M tokens`),
                 output: pricing(`$0.40 / 1M tokens`),
@@ -1474,7 +1474,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-4-1106-preview',
             modelName: 'gpt-4-1106-preview',
-            modelDescription: 'November 2023 preview version of GPT-4 Turbo with improved instruction following and a 128K token context window.',
+            modelDescription: 'November 2023 preview version of GPT-4 Turbo with 128K token context window. Features improved instruction following, better function calling capabilities, and enhanced reasoning. Includes knowledge cutoff from April 2023. Suitable for complex applications requiring extensive document understanding and sophisticated interactions.',
             pricing: {
                 prompt: pricing(`$10.00 / 1M tokens`),
                 output: pricing(`$30.00 / 1M tokens`),
@@ -1486,7 +1486,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-4-0125-preview',
             modelName: 'gpt-4-0125-preview',
-            modelDescription: 'January 2024 preview version of GPT-4 Turbo with improved reasoning capabilities and a 128K token context window.',
+            modelDescription: 'January 2024 preview version of GPT-4 Turbo with 128K token context window. Features improved reasoning capabilities, enhanced tool use, and more reliable function calling. Includes knowledge cutoff from October 2023. Offers better performance on complex logical tasks and more consistent outputs than previous preview versions.',
             pricing: {
                 prompt: pricing(`$10.00 / 1M tokens`),
                 output: pricing(`$30.00 / 1M tokens`),
@@ -1504,7 +1504,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-3.5-turbo-0125',
             modelName: 'gpt-3.5-turbo-0125',
-            modelDescription: 'January 2024 version of GPT-3.5 Turbo with improved reasoning capabilities and a 16K token context window.',
+            modelDescription: 'January 2024 version of GPT-3.5 Turbo with 16K token context window. Features improved reasoning capabilities, better instruction adherence, and reduced hallucinations compared to previous versions. Includes knowledge cutoff from September 2021. Provides good performance for most general applications at reasonable cost.',
             pricing: {
                 prompt: pricing(`$0.50 / 1M tokens`),
                 output: pricing(`$1.50  / 1M tokens`),
@@ -1516,7 +1516,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-4-turbo-preview',
             modelName: 'gpt-4-turbo-preview',
-            modelDescription: 'Preview version of GPT-4 Turbo that points to the latest model version. Features improved instruction following, 128K token context window and lower latency.',
+            modelDescription: 'Preview version of GPT-4 Turbo with 128K token context window that points to the latest development model. Features cutting-edge improvements to instruction following, knowledge representation, and tool use capabilities. Provides access to newest features but may have occasional behavior changes. Best for non-critical applications wanting latest capabilities.',
             pricing: {
                 prompt: pricing(`$10.00 / 1M tokens`),
                 output: pricing(`$30.00 / 1M tokens`),
@@ -1528,11 +1528,10 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'EMBEDDING',
             modelTitle: 'text-embedding-3-large',
             modelName: 'text-embedding-3-large',
-            modelDescription: "OpenAI's most capable text embedding model designed for high-quality embeddings for complex similarity tasks and information retrieval.",
+            modelDescription: "OpenAI's most capable text embedding model generating 3072-dimensional vectors. Designed for high-quality embeddings for complex similarity tasks, clustering, and information retrieval. Features enhanced cross-lingual capabilities and significantly improved performance on retrieval and classification benchmarks. Ideal for sophisticated RAG systems and semantic search applications.",
             pricing: {
                 prompt: pricing(`$0.13 / 1M tokens`),
-                // TODO: [🏏] Leverage the batch API @see https://platform.openai.com/docs/guides/batch
-                output: 0, // <- Note: [🆖] In Embedding models you dont pay for output
+                output: 0,
             },
         },
         /**/
@@ -1541,11 +1540,10 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'EMBEDDING',
             modelTitle: 'text-embedding-3-small',
             modelName: 'text-embedding-3-small',
-            modelDescription: 'Cost-effective embedding model with good performance for simpler tasks like text similarity and retrieval. Good balance of quality and efficiency.',
+            modelDescription: 'Cost-effective embedding model generating 1536-dimensional vectors. Balances quality and efficiency for simpler tasks while maintaining good performance on text similarity and retrieval applications. Offers 20% better quality than ada-002 at significantly lower cost. Ideal for production embedding applications with cost constraints.',
             pricing: {
                 prompt: pricing(`$0.02 / 1M tokens`),
-                // TODO: [🏏] Leverage the batch API @see https://platform.openai.com/docs/guides/batch
-                output: 0, // <- Note: [🆖] In Embedding models you dont pay for output
+                output: 0,
             },
         },
         /**/
@@ -1554,7 +1552,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-3.5-turbo-0613',
             modelName: 'gpt-3.5-turbo-0613',
-            modelDescription: 'June 2023 version of GPT-3.5 Turbo with function calling capabilities and a 4K token context window.',
+            modelDescription: "June 2023 version of GPT-3.5 Turbo with 4K token context window. Features function calling capabilities for structured data extraction and API interaction. Includes knowledge cutoff from September 2021. Maintained for applications specifically designed for this version's behaviors and capabilities.",
             pricing: {
                 prompt: pricing(`$1.50 / 1M tokens`),
                 output: pricing(`$2.00 / 1M tokens`),
@@ -1566,11 +1564,10 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'EMBEDDING',
             modelTitle: 'text-embedding-ada-002',
             modelName: 'text-embedding-ada-002',
-            modelDescription: 'Legacy text embedding model suitable for text similarity and retrieval augmented generation use cases. Replaced by newer embedding-3 models.',
+            modelDescription: 'Legacy text embedding model generating 1536-dimensional vectors suitable for text similarity and retrieval applications. Processes up to 8K tokens per request with consistent embedding quality. While superseded by newer embedding-3 models, still maintains adequate performance for many semantic search and classification tasks.',
             pricing: {
                 prompt: pricing(`$0.1 / 1M tokens`),
-                // TODO: [🏏] Leverage the batch API @see https://platform.openai.com/docs/guides/batch
-                output: 0, // <- Note: [🆖] In Embedding models you dont pay for output
+                output: 0,
             },
         },
         /**/
@@ -1597,7 +1594,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-4o-2024-05-13',
             modelName: 'gpt-4o-2024-05-13',
-            modelDescription: 'May 2024 version of GPT-4o with enhanced multimodal capabilities, improved reasoning, and optimized for vision, audio and chat at lower latencies.',
+            modelDescription: 'May 2024 version of GPT-4o with 128K context window. Features enhanced multimodal capabilities including superior image understanding (up to 20MP), audio processing, and improved reasoning. Optimized for 2x lower latency than GPT-4 Turbo while maintaining high performance. Includes knowledge up to October 2023. Ideal for production applications requiring reliable multimodal capabilities.',
             pricing: {
                 prompt: pricing(`$5.00 / 1M tokens`),
                 output: pricing(`$15.00 / 1M tokens`),
@@ -1609,7 +1606,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-4o',
             modelName: 'gpt-4o',
-            modelDescription: "OpenAI's most advanced multimodal model optimized for performance, speed, and cost. Capable of vision, reasoning, and high quality text generation.",
+            modelDescription: "OpenAI's most advanced general-purpose multimodal model with 128K context window. Optimized for balanced performance, speed, and cost with 2x faster responses than GPT-4 Turbo. Features excellent vision processing, audio understanding, reasoning, and text generation quality. Represents optimal balance of capability and efficiency for most advanced applications.",
             pricing: {
                 prompt: pricing(`$5.00 / 1M tokens`),
                 output: pricing(`$15.00 / 1M tokens`),
@@ -1621,7 +1618,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-4o-mini',
             modelName: 'gpt-4o-mini',
-            modelDescription: 'Smaller, more cost-effective version of GPT-4o with good performance across text, vision, and audio tasks at reduced complexity.',
+            modelDescription: 'Smaller, more cost-effective version of GPT-4o with 128K context window. Maintains impressive capabilities across text, vision, and audio tasks while operating at significantly lower cost. Features 3x faster inference than GPT-4o with good performance on general tasks. Excellent for applications requiring good quality multimodal capabilities at scale.',
             pricing: {
                 prompt: pricing(`$0.15 / 1M tokens`),
                 output: pricing(`$0.60 / 1M tokens`),
@@ -1633,7 +1630,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'o1-preview',
             modelName: 'o1-preview',
-            modelDescription: 'Advanced reasoning model with exceptional performance on complex logical, mathematical, and analytical tasks. Built for deep reasoning and specialized professional tasks.',
+            modelDescription: 'Advanced reasoning model with 128K context window specializing in complex logical, mathematical, and analytical tasks. Features exceptional step-by-step problem-solving capabilities, advanced mathematical and scientific reasoning, and superior performance on STEM-focused problems. Significantly outperforms GPT-4 on quantitative reasoning benchmarks. Ideal for professional and specialized applications.',
             pricing: {
                 prompt: pricing(`$15.00 / 1M tokens`),
                 output: pricing(`$60.00 / 1M tokens`),
@@ -1645,8 +1642,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'o1-preview-2024-09-12',
             modelName: 'o1-preview-2024-09-12',
-            modelDescription: 'September 2024 version of O1 preview with specialized reasoning capabilities for complex tasks requiring precise analytical thinking.',
-            //             <- TODO: [💩] Some better system to organize these date suffixes and versions
+            modelDescription: 'September 2024 version of O1 preview with 128K context window. Features specialized reasoning capabilities with 30% improvement on mathematical and scientific accuracy over previous versions. Includes enhanced support for formal logic, statistical analysis, and technical domains. Optimized for professional applications requiring precise analytical thinking and rigorous methodologies.',
             pricing: {
                 prompt: pricing(`$15.00 / 1M tokens`),
                 output: pricing(`$60.00 / 1M tokens`),
@@ -1658,7 +1654,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'o1-mini',
             modelName: 'o1-mini',
-            modelDescription: 'Smaller, cost-effective version of the O1 model with good performance on reasoning tasks while maintaining efficiency for everyday analytical use.',
+            modelDescription: 'Smaller, cost-effective version of the O1 model with 128K context window. Maintains strong analytical reasoning abilities while reducing computational requirements by 70%. Features good performance on mathematical, logical, and scientific tasks at significantly lower cost than full O1. Excellent for everyday analytical applications that benefit from reasoning focus.',
             pricing: {
                 prompt: pricing(`$3.00 / 1M tokens`),
                 output: pricing(`$12.00 / 1M tokens`),
@@ -1670,7 +1666,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'o1',
             modelName: 'o1',
-            modelDescription: "OpenAI's advanced reasoning model focused on logic and problem-solving. Designed for complex analytical tasks with rigorous step-by-step reasoning. 128K context window.",
+            modelDescription: "OpenAI's advanced reasoning model with 128K context window focusing on logical problem-solving and analytical thinking. Features exceptional performance on quantitative tasks, step-by-step deduction, and complex technical problems. Maintains 95%+ of o1-preview capabilities with production-ready stability. Ideal for scientific computing, financial analysis, and professional applications.",
             pricing: {
                 prompt: pricing(`$15.00 / 1M tokens`),
                 output: pricing(`$60.00 / 1M tokens`),
@@ -1682,11 +1678,10 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'o3-mini',
             modelName: 'o3-mini',
-            modelDescription: 'Cost-effective reasoning model optimized for academic and scientific problem-solving. Efficient performance on STEM tasks with deep mathematical and scientific knowledge. 128K context window.',
+            modelDescription: 'Cost-effective reasoning model with 128K context window optimized for academic and scientific problem-solving. Features efficient performance on STEM tasks with specialized capabilities in mathematics, physics, chemistry, and computer science. Offers 80% of O1 performance on technical domains at significantly lower cost. Ideal for educational applications and research support.',
             pricing: {
                 prompt: pricing(`$3.00 / 1M tokens`),
                 output: pricing(`$12.00 / 1M tokens`),
-                // <- TODO: !! Unsure, check the pricing
             },
         },
         /**/
@@ -1695,7 +1690,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'o1-mini-2024-09-12',
             modelName: 'o1-mini-2024-09-12',
-            modelDescription: "September 2024 version of O1-mini with balanced reasoning capabilities and cost-efficiency. Good for analytical tasks that don't require the full O1 model.",
+            modelDescription: "September 2024 version of O1-mini with 128K context window featuring balanced reasoning capabilities and cost-efficiency. Includes 25% improvement in mathematical accuracy and enhanced performance on coding tasks compared to previous versions. Maintains efficient resource utilization while delivering improved results for analytical applications that don't require the full O1 model.",
             pricing: {
                 prompt: pricing(`$3.00 / 1M tokens`),
                 output: pricing(`$12.00 / 1M tokens`),
@@ -1707,7 +1702,7 @@ const OPENAI_MODELS = exportJson({
             modelVariant: 'CHAT',
             modelTitle: 'gpt-3.5-turbo-16k-0613',
             modelName: 'gpt-3.5-turbo-16k-0613',
-            modelDescription: 'June 2023 version of GPT-3.5 Turbo with extended 16k token context window for processing longer conversations and documents.',
+            modelDescription: "June 2023 version of GPT-3.5 Turbo with extended 16K token context window. Features good handling of longer conversations and documents with improved memory management across extended contexts. Includes knowledge cutoff from September 2021. Maintained for applications specifically designed for this version's behaviors and capabilities.",
             pricing: {
                 prompt: pricing(`$3.00 / 1M tokens`),
                 output: pricing(`$4.00 / 1M tokens`),
@@ -2242,7 +2237,7 @@ class OpenAiAssistantExecutionTools extends OpenAiExecutionTools {
         });
         const rawRequest = {
             // TODO: [👨‍👨‍👧‍👧] ...modelSettings,
-            // TODO: [👨‍👨‍👧‍👧][🧠] What about system message for assistants, does it make sence - combination of OpenAI assistants with Promptbook Personas
+            // TODO: [👨‍👨‍👧‍👧][🧠] What about system message for assistants, does it make sense - combination of OpenAI assistants with Promptbook Personas
             assistant_id: this.assistantId,
             thread: {
                 messages: [
@@ -2330,7 +2325,7 @@ class OpenAiAssistantExecutionTools extends OpenAiExecutionTools {
     }
 }
 /**
- * TODO: [🧠][🧙‍♂️] Maybe there can be some wizzard for thoose who want to use just OpenAI
+ * TODO: [🧠][🧙‍♂️] Maybe there can be some wizard for those who want to use just OpenAI
  * TODO: Maybe make custom OpenAiError
  * TODO: [🧠][🈁] Maybe use `isDeterministic` from options
  * TODO: [🧠][🌰] Allow to pass `title` for tracking purposes
@@ -2523,7 +2518,7 @@ const $llmToolsRegister = new $Register('llm_execution_tools_constructors');
  * Warning: This is not useful for the end user, it is just a side effect of the mechanism that handles all available LLM tools
  *
  * @public exported from `@promptbook/openai`
- * @public exported from `@promptbook/wizzard`
+ * @public exported from `@promptbook/wizard`
  * @public exported from `@promptbook/cli`
  */
 const _OpenAiRegistration = $llmToolsRegister.register(createOpenAiExecutionTools);
@@ -2533,7 +2528,7 @@ const _OpenAiRegistration = $llmToolsRegister.register(createOpenAiExecutionTool
  * Note: [🏐] Configurations registrations are done in register-constructor.ts BUT constructor register-constructor.ts
  *
  * @public exported from `@promptbook/openai`
- * @public exported from `@promptbook/wizzard`
+ * @public exported from `@promptbook/wizard`
  * @public exported from `@promptbook/cli`
  */
 const _OpenAiAssistantRegistration = $llmToolsRegister.register(createOpenAiAssistantExecutionTools);