mixseek-model-list

MixSeek Model List

Overview

Provides the list of LLM models available in MixSeek-Core. It dynamically retrieves the latest model information via the API, allowing you to check model information by provider, recommended settings by use case, and compatibility information with agent_type.

FR-008 compliant: Retrieve the model list via the API from each provider: Google Gemini, Anthropic Claude, OpenAI, and Grok.

Article 6 compliant: If an API key is not set or the API fails, an error is explicitly reported (no implicit fallback).

Prerequisites

Environment variables (required for API retrieval)

To retrieve the model list from the API, set the corresponding environment variables:

Provider	Environment variable	Notes
Google	`GOOGLE_API_KEY`
Anthropic	`ANTHROPIC_API_KEY`
OpenAI	`OPENAI_API_KEY`
Grok	`GROK_API_KEY`
Groq	`GROQ_API_KEY`	`gsk_` Starts with (mixseek-plus extension)
Tavily	`TAVILY_API_KEY`	`tvly-` Starts with. Required for web_search-type operations (optional)
ClaudeCode	-	No environment variable required due to CLI authentication (mixseek-plus extension)

Note: If an environment variable is not set, the corresponding provider explicitly reports an error.

Usage

Step 1: Check environment variables

Check whether the API key is set:

# 設定状況を確認
echo "GOOGLE_API_KEY: ${GOOGLE_API_KEY:+設定済み}"
echo "ANTHROPIC_API_KEY: ${ANTHROPIC_API_KEY:+設定済み}"
echo "OPENAI_API_KEY: ${OPENAI_API_KEY:+設定済み}"
echo "GROK_API_KEY: ${GROK_API_KEY:+設定済み}"

Step 2: Check requirements

Check the user's use case:

All models list: Display models from all providers
Provider specification: Display only models from a specific provider
Recommendations by use case: Models suitable for use cases such as Leader/Member/Evaluator
API retrieval: Retrieve the latest model information from the API

Step 3: Retrieve model information

API retrieval via script (recommended):

# 全プロバイダーからモデル取得（MixSeek形式）
skills/detect-python-command/scripts/run-python.sh \
    skills/mixseek-model-list/scripts/fetch-models.py

# 特定プロバイダーのみ
skills/detect-python-command/scripts/run-python.sh \
    skills/mixseek-model-list/scripts/fetch-models.py --provider google

# JSON形式で出力
skills/detect-python-command/scripts/run-python.sh \
    skills/mixseek-model-list/scripts/fetch-models.py --json

# 詳細出力（エラー情報も表示）
skills/detect-python-command/scripts/run-python.sh \
    skills/mixseek-model-list/scripts/fetch-models.py --verbose --format text

CLI Command Reference

fetch-models.py [OPTIONS]

OPTIONS:
  --provider TEXT    プロバイダー指定: google, anthropic, openai, grok, all
                     (default: all)
  --format TEXT      出力形式: text, json, csv, mixseek
                     (default: mixseek)
  --json             --format json のショートカット
  --verbose, -v      詳細出力（エラー情報も表示）

Output format

Format	Description	Use
`mixseek`	`provider:model-id` format (one model per line)	Direct copy to team configuration
`json`	JSON format (including metadata)	Program integration
`text`	Detailed display by provider	For human reading
`csv`	CSV format	Spreadsheets, etc.

Model format

In MixSeek-Core, models are specified in the following format:

provider:model-name

Provider	Prefix	Environment variable	Representative model
Google	`google-gla`	`GOOGLE_API_KEY`	`gemini-2.5-pro`, `gemini-2.5-flash`
Anthropic	`anthropic`	`ANTHROPIC_API_KEY`	`claude-sonnet-4-5`, `claude-haiku-4-5`
OpenAI	`openai`	`OPENAI_API_KEY`	`gpt-4o`, `gpt-4o-mini`
Grok	`grok`	`GROK_API_KEY`	`grok-3`, `grok-3-mini`
Groq	`groq`	`GROQ_API_KEY`	`llama-3.3-70b-versatile`, `qwen/qwen3-32b`
ClaudeCode	`claudecode`	- (CLI authentication)	`claude-sonnet-4-5`, `claude-haiku-4-5`

Note: Groq と ClaudeCode は mixseek-plus extended provider. It cannot be used with mixseek-core alone.

Retrieve the list of available models with the fetch-models.py script.

Provider Characteristics

Provider	Features
Google	High quality and stable. For Leader Agent / Evaluator
Anthropic	`code_execution` Fully compatible with agent_type
OpenAI	Stable and highly versatile
Grok	Some models have built-in web search functionality
Groq	Low latency and high-speed inference. For real-time responses (mixseek-plus extension)
ClaudeCode	Integrated built-in tools (Bash, file operations, web search). For development tasks (mixseek-plus extension)

Recommended by Use Case

Leader Agent

For leaders who coordinate and direct tasks. High-quality (-proseries) models are recommended.

Member Agent

For members responsible for task execution. Fast and cost-efficient (-flash、-miniseries) is recommended.

code_execution Agent

When code execution is required.Only compatible with Anthropic models。

Evaluator / Judgment

For evaluation and judgment. Stable and consistent high-quality models are recommended.

Real-time responses (mixseek-plus extension)

When low latency is required.Groq modelsis recommended.groq:llama-3.3-70b-versatile etc.

Development task automation (mixseek-plus extension)

Development tasks combining file operations, command execution, and web search.ClaudeCode modelsis recommended. Automate complex workflows with built-in tools.

agent_type Compatibility

mixseek-core standard types

agent_type	Google	Anthropic	OpenAI	Grok
`plain`	✓	✓	✓	✓
`web_search`	✓	✓	✓	✓
`code_execution`	✗	✓	✗	✗
`web_fetch`	✓	✓	✓	✓
`custom`	✓	✓	✓	✓

Note: code_executionis only compatible with Anthropic models

mixseek-plus extension types

agent_type	Groq	ClaudeCode	Description
`groq_plain`	✓	-	Groq basic agent. High-speed inference
`groq_web_search`	✓	-	Groq + Tavily search integration
`tavily_search`	✓	✓	Specialized in Tavily API search
`claudecode_plain`	-	✓	ClaudeCode basic agent
`claudecode_tavily_search`	-	✓	ClaudeCode + Tavily search integration
`custom`	✓	✓	Custom tool integration

Note: Extension types are available only in mixseek-plus .

Example

Retrieving the full model list

# スクリプトを実行してAPIから最新のモデル一覧を取得
skills/detect-python-command/scripts/run-python.sh \
    skills/mixseek-model-list/scripts/fetch-models.py --format text

Example output:

[GOOGLE]
  Prefix: google-gla
  Models:
    - gemini-2.5-pro: Gemini 2.5 Pro
    - gemini-2.5-flash: Gemini 2.5 Flash
    ...

[ANTHROPIC]
  Prefix: anthropic
  Models:
    - claude-sonnet-4-5-...: Claude Sonnet 4.5 [code_exec]
    ...

Checking models compatible with code_execution

# JSON形式で取得し、code_exec_compatible を確認
skills/detect-python-command/scripts/run-python.sh \
    skills/mixseek-model-list/scripts/fetch-models.py --provider anthropic --json

Point: code_exec_compatible: true models in code_execution compatible with agent_type

Example usage in team settings

[[team.members]]
agent_name = "coder"
agent_type = "code_execution"
model = "anthropic:<model-id>"  # fetch-models.py で取得したモデルIDを使用

Troubleshooting

API key error (fetch-models.py)

Error [google]: Error: GOOGLE_API_KEY environment variable is not set.

Solution: Set the corresponding environment variable:

export GOOGLE_API_KEY="your-api-key"
export ANTHROPIC_API_KEY="your-api-key"
export OPENAI_API_KEY="your-api-key"
export GROK_API_KEY="your-api-key"

API connection error

Error [google]: <urlopen error ...>

Cause and solution:

Check network connection
Check proxy settings (in corporate environments)
Check the validity of the API key
If other providers succeed, only that model will be output

API authentication error (401/403)

Warning: API fetch failed for openai: HTTP Error 401: Unauthorized

Solution:

Check that the API key is set correctly
Check the expiration date of the API key
Check the permissions (scope) of the API key

Timeout error

Error [anthropic]: timed out

Solution:

Check network connection
Retry later
Specify only a specific provider: --provider openai

API key error when configuring MixSeek

Error: API key not found for provider: google-gla

Solution: Set the corresponding environment variable, then run MixSeek

Model not found

Error: Unknown model: invalid-model

Solution:

provider:model-name Check the format
Use a valid model name (check with this skill)

code_execution unsupported error

Error: code_execution not supported for model: google-gla:...

Solution:

Switch to an Anthropic model (code_executionsupports Anthropic only)
fetch-models.py --provider anthropic Check supported models with

mixseek-plus extension

This section is mixseek-plus explains its unique features. They are not available with mixseek-core alone.

Extended providers

Provider	Features	Main use cases
Groq	Low latency and high-speed inference. Accelerated by LPU (Language Processing Unit)	Real-time responses, large-scale batch processing
ClaudeCode	Built-in tool integration. Includes Bash execution, file operations, and web search as standard	Development task automation, codebase exploration

Extended agent_type

agent_type	Provider	Description
`groq_plain`	Groq	Basic dialogue. When high-speed inference is required
`groq_web_search`	Groq	Web search integration using the Tavily API
`tavily_search`	Groq / ClaudeCode	Tavily search specialized agent
`claudecode_plain`	ClaudeCode	Basic dialogue. Built-in tools available
`claudecode_tavily_search`	ClaudeCode	ClaudeCode + Tavily search integration

Team configuration example (mixseek-plus)

# Groqエージェント（高速応答）
[[team.members]]
agent_name = "fast_responder"
agent_type = "groq_plain"
model = "groq:llama-3.3-70b-versatile"

# ClaudeCodeエージェント（開発タスク）
[[team.members]]
agent_name = "developer"
agent_type = "claudecode_plain"
model = "claudecode:claude-sonnet-4-5"

# Tavily検索エージェント
[[team.members]]
agent_name = "researcher"
agent_type = "tavily_search"
model = "groq:llama-3.3-70b-versatile"

Notes

mixseek-plus dependencies: To use these features, the mixseek-plus package is required
Environment variables: Groqは GROQ_API_KEY、Tavilyは TAVILY_API_KEY is required
ClaudeCode authentication: Uses Claude CLI authentication. No environment variables are required, but theclaude command must be authenticated
Compatibility: Not compatible with mixseek-core standard agent_type. Use the provider dedicated to extended types

References

Script: scripts/fetch-models.py
Team configuration: skills/mixseek-team-config/
Evaluation configuration: skills/mixseek-evaluator-config/

About this skill

Overview

Prerequisites

Environment variables (required for API retrieval)

Usage

Step 1: Check environment variables

Step 2: Check requirements

Step 3: Retrieve model information

CLI Command Reference

Output format

Model format

Provider Characteristics

Recommended by Use Case

Leader Agent

Member Agent

code_execution Agent

Evaluator / Judgment

Real-time responses (mixseek-plus extension)

Development task automation (mixseek-plus extension)

agent_type Compatibility

mixseek-core standard types

mixseek-plus extension types

Example

Retrieving the full model list

Checking models compatible with code_execution

Example usage in team settings

Troubleshooting

API key error (fetch-models.py)

API connection error

API authentication error (401/403)

Timeout error

API key error when configuring MixSeek

Model not found

code_execution unsupported error

mixseek-plus extension

Extended providers

Extended agent_type

Team configuration example (mixseek-plus)

Notes

References