Skill

tooyoung:easy-openrouter

Tests LLM models via OpenRouter using bash script, measuring latency, cost, token usage, and outputs. Supports model ID formats, :nitro/:online modifiers, rankings, and provider comparisons for optimal selection.

Bash

OpenAI

Anthropic

ai-ml

npx claudepluginhub shiqkuangsan/oh-my-daily-skills

Tool Access

This skill uses the workspace's default tool permissions.

Preview

Quickly test LLM models and find the best one for your task.

Supporting Assets

scripts/call_openrouter.sh

SKILL.md

Similar Skills

openrouter-models

125

Queries OpenRouter for 300+ AI models' pricing, context lengths, capabilities, throughput, provider latency/uptime. Scripts to list, search, compare, resolve names, find optimal providers.

9 files

openrouter

openrouter-model-catalog

1.9k

Queries OpenRouter's 400+ LLM catalog API to list, filter by pricing/context/tools, and compare models/providers. Use for model selection in AI apps.

5 files5 tools

openrouter-pack

model-comparator

Compares AI and LLM models on benchmarks, capabilities, cost, latency, context window, and task-specific fit to help select optimal models for production use cases and budgets.

nickcrew-claude-ctx-plugin

Stats

Stars15

Forks1

Last CommitMay 1, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

Easy OpenRouter

Quickly test LLM models and find the best one for your task.

Prerequisites

export OPENROUTER_API_KEY="sk-or-v1-xxxx"

Quick Model Testing

Use scripts/call_openrouter.sh:

# Basic test
bash scripts/call_openrouter.sh \
  --model "anthropic/claude-sonnet-4" \
  --prompt "Explain quantum computing in one sentence" \
  --json

# With parameters
bash scripts/call_openrouter.sh \
  --model "openai/gpt-4o" \
  --prompt "Write a haiku about code" \
  --max-tokens 100 \
  --temperature 0.7 \
  --json

Output includes: response time (ms), cost ($), token usage, full response

Parameters:

--model: Model ID (required)
--prompt: Prompt text (required)
--system: System prompt (optional)
--max-tokens: Max tokens
--temperature: Temperature 0.0-2.0
--json: JSON output

Model Naming Format

provider/model-name[:modifier]

Common models:

anthropic/claude-sonnet-4
openai/gpt-4o
openai/gpt-4o-mini
google/gemini-2.0-flash-001
deepseek/deepseek-chat

Modifiers:

:nitro - Use fastest provider
:online - Enable web search
Combinable: anthropic/claude-sonnet-4:nitro:online

Find Optimal Models

Find Best Model by Category (Rankings)

https://openrouter.ai/rankings?category=<category>#categories

Category	URL
Programming	`https://openrouter.ai/rankings?category=programming#categories`
Translation	`https://openrouter.ai/rankings?category=translation#categories`
Science	`https://openrouter.ai/rankings?category=science#categories`
Finance	`https://openrouter.ai/rankings?category=finance#categories`
Legal	`https://openrouter.ai/rankings?category=legal#categories`
Academia	`https://openrouter.ai/rankings?category=academia#categories`
Marketing	`https://openrouter.ai/rankings?category=marketing#categories`
SEO	`https://openrouter.ai/rankings?category=marketing/seo#categories`

Find Provider by Performance/Price

https://openrouter.ai/<model-slug>/providers?sort=<sort>

Sort options:

throughput - By throughput (tokens/sec)
latency - By latency
price - By price

Example: Find fastest provider for claude-sonnet-4

https://openrouter.ai/anthropic/claude-sonnet-4/providers?sort=throughput

Get Full Model List

curl -s https://openrouter.ai/api/v1/models \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  | jq '.data | sort_by(.pricing.prompt) | .[:10] | .[].id'
# Output: top 10 cheapest models

Typical Workflows

1. Quick Model Comparison

# Test Claude
bash scripts/call_openrouter.sh --model "anthropic/claude-sonnet-4" --prompt "Explain microservices architecture" --json > claude.json

# Test GPT-4o
bash scripts/call_openrouter.sh --model "openai/gpt-4o" --prompt "Explain microservices architecture" --json > gpt4o.json

# Compare results
jq -s '[.[0], .[1]] | map({model, response_time_ms, cost})' claude.json gpt4o.json

2. Find Best Model for Programming Tasks

Open https://openrouter.ai/rankings?category=programming#categories
Note the top-ranked models
Test actual performance with the script

3. Choose Model for Chainlit Demo

Considerations:

For demos: Prioritize speed → use :nitro modifier
Cost-sensitive: Use gpt-4o-mini or deepseek-chat
Quality-first: Use claude-sonnet-4 or gpt-4o

Advanced

For API details, query OpenRouter documentation via Context7.