models.dev cli

Multi-column sorting CLI for models using the @opencode-ai/models package.

Why This Exists

The models.dev website only allows single-column sorting and combines input/output token costs into one column. This CLI lets you:

• Sort by multiple columns in priority order
• Separate input and output token costs for comparison
• Filter for reasoning models and compare their pricing
• Export data as JSON for further analysis

Installation

cd models-dev-cli
bun install

Requires Bun (v1.0+)

Usage

bun run models-dev-cli.ts [options]

Options

Flag	Description
--sort <field>[,asc|desc]	Sort by field (can specify multiple, applied left-to-right)
--limit <n>	Limit output to n rows (default: 50)
--show-reasoning	Show reasoning model details
--json	Output as JSON instead of formatted table
--help	Show help

Sortable Fields

• output_cost - Output token cost (USD per 1M tokens)
• input_cost - Input token cost (USD per 1M tokens)
• context - Context window size
• output_limit - Max output tokens
• reasoning - Reasoning support (true/false)
• tool_call - Tool/function calling support
• structured_output - Structured output (JSON schema) support
• release_date - Release date
• name, id - Model name/ID

Examples

Cheapest output cost (cheapest first):

bun run models-dev-cli.ts --sort output_cost,asc --limit 10

Most expensive models:

bun run models-dev-cli.ts --sort output_cost,desc --limit 10

Multi-column: Reasoning models first, then by price:

bun run models-dev-cli.ts --sort reasoning,desc --sort output_cost,asc --show-reasoning

Best context windows, sorted by price:

bun run models-dev-cli.ts --sort context,desc --sort output_cost --limit 20

Export JSON for analysis:

bun run models-dev-cli.ts --json --sort output_cost | jq '.[0:5]'

All reasoning models with pricing sorted:

bun run models-dev-cli.ts --sort reasoning,desc --sort output_cost --show-reasoning

Output Columns

Column	Meaning
Name	Human-readable model name
Output $	Cost per 1M output tokens (USD)
Input $	Cost per 1M input tokens (USD)
Context	Context window size
Max Out	Maximum output tokens
Rsn	Reasoning enabled (Yes/-)
Tool	Tool/function calling supported
Struct	Structured output (JSON schema) supported

Sample Output

Models (sorted by: output_cost↑, context↓)
Total: 20

Name                                     Output $       Input $     Context   Max Out  Rsn     Tool   Struct
------------------------------------------------------------------------------------------------------------
Voxtral Mini 3B 2507                        $0.04         $0.04      128.0K      4.1K  -          ✓        ✓
Gemma 3 4B IT                               $0.08         $0.04      128.0K      4.1K  -          ✓        -
Nova Micro                                  $0.14         $0.04      128.0K      8.2K  -          ✓        -
...

Data Source

Uses the bundled snapshot from @opencode-ai/models/snapshot - no API key required. Data is updated periodically in the npm package.

License

MIT