Read the LMArena coding LLM leaderboard

Ask for the Code Arena coding leaderboard without providing a model name. Compare the returned model names, positions, scores and vote counts to choose candidates for your own tasks.

Your AI

Use with your AI

For tasks with several steps, continue in the AI you already use.

CodexClaude CodeCursorWindsurfCline

View instructions

Use Toolmu get-code-leaderboard-lmarena to read Code Arena coding rankings. Reuse a working Toolmu connection, or follow https://toolmu.com/SKILL.md to connect. Check current availability, requirements and the Credits ceiling, then execute within my authorization and budget. No named input is documented; handle the request internally without asking me for model references, field names or empty objects, and do not invent filtering, year, count or pagination controls. Apply any ordinary display-count preference only to the actual returned records. Present coding records in source order with readable model names, ranks, scores, vote counts and relevant actual source links or organization labels. Keep actual source identities internally and distinguish similarly named variants without merging their records. Report coding results only; do not add an unreturned text snapshot or separate model list. Preserve exact source ranking-range labels and values, even if their numerical order is reversed, and explain missing information plainly. Do not infer accuracy, unique voters, confidence percentage, source update dates, price currency, context units or license rights. These records provide candidates for evaluation, not proof of the best model or a test of my project. Bracketed optional example text is replaced by my own content.

Task:
Show the LMArena coding leaderboard with the model names, ranks, scores and vote counts returned. Keep missing information clear and explain the limits of using these rankings to choose coding models.

Paste into your connected AI's conversation.

Connection guide

Up to 0.012 credits / call

The final charge is based on confirmed usage. Unused reserved credits return to your available balance.

LMArena

Read the LMArena coding LLM leaderboard

Ask for the Code Arena coding leaderboard without providing a model name. Compare the returned model names, positions, scores and vote counts to choose candidates for your own tasks.

You receive

Code Arena coding-model records with model names, ranks, scores, vote counts and any available source model information. Read the returned coding records in source order.

Start with your task

Show the LMArena coding leaderboard with the model names, ranks, scores and vote counts returned. Keep missing information clear and explain the limits of using these rankings to choose coding models.

What to know

This retrieves coding rankings without running your tasks or testing your codebase. The response does not specify ranking-update time, rating method, vote-counting rules, confidence level, context units or price currency. Keep the original rank-bound labels and unknown metadata; these records do not establish a universally best model or current vendor terms.

Inputs & outputs

Inspect the current inputs, availability, and maximum charge before starting.

Input schema

Input schema
Field / Path	Type	Requirement	Meaning
Optional request query container`queryParams`	`object`	Optional	Optional open object, without named properties or documented defaults. Request the leaderboard with empty input; the schema alone does not establish search, filtering, a year, pagination or a record limit.

View raw JSON

{
  "type": "object",
  "required": [],
  "properties": {
    "queryParams": {
      "type": "object",
      "properties": {},
      "additionalProperties": true
    }
  },
  "additionalProperties": false
}

Result schema

No fixed output contract is available. Field types reflect saved result evidence; individual responses can differ.

Result schema
Field / Path	Type	Meaning
Coding-model entries`result.data.entries`	`array`	Source coding records. The saved response contains 138 entries; this observed size does not guarantee a fixed future count or coverage of every model. No separate text or model-list collection is shown.
Source model identifier`result.data.entries[].modelKey`	`string`	The exact source key for this entry. Preserve it across comparisons; similar display names do not establish identical model configurations or interchangeable records.
Model display name`result.data.entries[].modelDisplayName`	`string`	Name reported by the source, including its version or configuration wording. Keep that wording with modelKey instead of shortening away distinctions.
Reported organization`result.data.entries[].modelOrganization`	`string`	Source organization text. Two saved records contain an empty string; leave that blank rather than substituting an inferred owner.
Source model reference`result.data.entries[].modelUrl`	`string`	Reference URL supplied for the model. The response does not verify that it grants executable access or states the current specifications, prices or terms.
Coding position`result.data.entries[].rank`	`number`	Integer position returned in this coding collection. It describes the source ordering, without guaranteeing correctness on your project or performance in other task categories.
Reported coding rating`result.data.entries[].rating`	`number`	Numerical source rating with no documented scale or calculation here. It is not a percentage accuracy, pass rate or probability of task success.
Reported vote count`result.data.entries[].votes`	`number`	Source integer count. The counting rules and population are unspecified; it cannot be assumed to mean distinct people or completed coding tasks.
Source rankLower`result.data.entries[].rankLower`	`number`	Literal integer source field. It exceeds rankUpper in 136 of the 138 saved entries; do not swap the labels or infer a conventional confidence interval.
Source rankUpper`result.data.entries[].rankUpper`	`number`	Literal paired source value. Preserve it under this exact name even if it is numerically smaller than rankLower; the statistical convention and coverage are unknown.
Lower rating value`result.data.entries[].ratingLower`	`number`	Source numerical bound, at or below rating in the saved records. Its estimation method and confidence level are not defined by this endpoint.
Upper rating value`result.data.entries[].ratingUpper`	`number`	Source numerical bound, at or above the saved rating. Read it as reported without adding a significance test or confidence percentage.
Reported license name`result.data.entries[].license`	`string`	Source license string for reference. Check the applicable license and intended use separately; this label does not verify commercial, redistribution or other permission.
Reported release label`result.data.entries[].releaseType`	`null \| string`	Source string or null, with 131 saved values null. No label definitions are documented, and a null value does not establish whether the model is available.
Reported context value`result.data.entries[].contextLength`	`number \| null`	Integer or null; 20 saved records are null. The unit and supported conditions are unspecified, so this is not a verified context-capacity guarantee.
Reported input price`result.data.entries[].inputPricePerMillion`	`number \| null`	Number or null; eight saved values are null. Although its name says per million, the currency, priced quantity and conditions are not defined. Unknown prices remain null rather than zero.
Reported output price`result.data.entries[].outputPricePerMillion`	`number \| null`	Number or null, separate from the input price; eight saved values are null. It is source metadata with unspecified currency and terms, rather than a verified current quote.

Additional notes

Use this coding leaderboard when you want a source-based starting point for choosing models to evaluate. Read entries in the returned order and retain each modelKey with its display name and configuration wording. The current input definition has no named required control; an empty request is sufficient to request the collection. The open query container does not document model search, year, filter, pagination or top-N options.

Read rank, rating and votes together rather than treating a single position as evidence of correctness on your own tasks. Preserve rankLower and rankUpper exactly: in most saved entries their numerical order differs from a conventional ascending interval. The source does not define the statistical method or confidence level, so these fields should not become a confidence percentage or an accuracy claim.

Use modelUrl and other reported metadata as references for further evaluation. Some context, release and price values are null, and an organization label can be blank. Retain those gaps; check current model documentation, pricing and license terms before making an implementation decision. This response contains coding entries, without a separate text leaderboard or a source ranking-update timestamp.

Questions about this tool.

Inputs, results and details to know before running it.

Get connected

It reports source coding positions, ratings and votes. It does not execute your prompts, inspect your repository or measure task-specific correctness. Use it to choose candidates, then run an evaluation suited to your work.

Read the LMArena coding LLM leaderboard

Read the LMArena coding LLM leaderboard

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Ask for the coding leaderboard

Read the returned models

Choose candidates to evaluate

Questions about this tool.

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Ask for the coding leaderboard

Read the returned models

Choose candidates to evaluate

Questions about this tool.

Does this ranking show which model will solve my tasks?

How should I read an unclear ranking range?

Are the price and license labels ready for a deployment decision?