Review LMArena coding rankings with text context

Ask for LMArena’s coding-model rankings without providing a model name. Read the coding results and the separate text-task ranking snapshot to prepare a list of models to evaluate.

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-best-coding-models-lmarena to read coding rankings and the separate text-ranking snapshot. 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. This documented operation requires no extra input from me; handle the request internally without inventing model, year, filtering, count or pagination controls. A requested display count is applied to actual returned records, not sent as an unsupported source filter. Present coding and text records separately, in their source order, with readable model names, ranks, scores, vote counts and available source labels or links. Retain each actual source identity internally and distinguish variants by readable context rather than merging by name. Do not compare the two score sets on an assumed common scale. Preserve exact source rank-bound labels and values, including reversed numerical order, and explain missing information plainly. Do not infer confidence level, accuracy, unique voters, source update dates, model price currency, context units or license permissions. These rankings provide candidates to test, not a universally best model or a test of my project. Bracketed optional example text is replaced by my own content.

Task:
Show LMArena’s coding-model rankings with the model names, ranks, scores and vote counts returned. Present the text-task rankings separately as background, and explain any missing information or limits before I choose models to evaluate.

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

Review LMArena coding rankings with text context

Ask for LMArena’s coding-model rankings without providing a model name. Read the coding results and the separate text-task ranking snapshot to prepare a list of models to evaluate.

You receive

Two separate sets of LMArena records: coding-model rankings and a text-model ranking snapshot. Read the returned model names, ranks, scores, vote counts and available model information.

Start with your task

Show LMArena’s coding-model rankings with the model names, ranks, scores and vote counts returned. Present the text-task rankings separately as background, and explain any missing information or limits before I choose models to evaluate.

What to know

Source rankings do not test your codebase or establish a universally best model. Keep coding and text scores separate. The response has no leaderboard-update time and does not define rating methodology, vote-counting rules, confidence level, context units or model-price currency. Preserve source rank-bound labels as returned; null prices are unknown.

Inputs & outputs

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

Input schema

Input schema
Field / Path	Type	Requirement	Meaning
Optional query container`queryParams`	`object`	Optional	Optional object with no documented named controls. Read the combined snapshot with an empty input; openness does not establish filter, model, year, top-N or pagination support.

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
Code Arena records`result.data.code_leaderboard`	`array`	Coding collection as reported by the source. The whole saved response has 138 entries; that observed count is not a fixed future size or a complete model-universe guarantee.
Coding source model key`result.data.code_leaderboard[].modelKey`	`string`	Original source identifier. Different keys must stay distinct, even when display names match; the identifier is not itself a promise of an executable model endpoint.
Coding model display name`result.data.code_leaderboard[].modelDisplayName`	`string`	Original source name, including any version or configuration wording. Preserve it rather than assuming models with similar names are equivalent.
Coding organization label`result.data.code_leaderboard[].modelOrganization`	`string`	Source organization string. Two saved coding entries have an empty label; retain the blank instead of inventing an organization or verified ownership.
Coding model reference link`result.data.code_leaderboard[].modelUrl`	`string`	Source-provided model URL. It may point to documentation or a model page; access, execution, current specifications and pricing are not verified by the leaderboard response.
Source coding rank`result.data.code_leaderboard[].rank`	`number`	Integer source position within the coding collection. Keep it with that collection; it is not a text rank or a measured guarantee for your project.
Source coding rating`result.data.code_leaderboard[].rating`	`number`	Numerical source rating. The endpoint does not define its scale or computation; do not convert it to percentage accuracy, pass rate or success probability.
Source coding vote count`result.data.code_leaderboard[].votes`	`number`	Integer count reported by the source. Counting rules, population and independence are not specified; do not label it unique users or successful coding tasks.
Original rankLower value`result.data.code_leaderboard[].rankLower`	`number`	Keep the source field name and value. It exceeds rankUpper in 136 of 138 saved coding rows; the source convention and confidence coverage are not defined here.
Original rankUpper value`result.data.code_leaderboard[].rankUpper`	`number`	Keep this source value beside rankLower without reversing or relabeling the fields. Do not treat the two values as a conventional ascending interval.
Source lower rating value`result.data.code_leaderboard[].ratingLower`	`number`	Source numerical rating bound. It is at or below the saved rating, but the endpoint does not specify estimation method or confidence level.
Source upper rating value`result.data.code_leaderboard[].ratingUpper`	`number`	Source numerical rating bound. It is at or above the saved rating; preserve it without inventing a statistical significance or confidence claim.
Source license label`result.data.code_leaderboard[].license`	`string`	Original license string. It is a reference label, not verification of commercial use, redistribution or other rights; inspect the actual applicable license for those decisions.
Source release-type label`result.data.code_leaderboard[].releaseType`	`null \| string`	Source string or null; 131 saved coding entries are null. No meaning for its labels is documented here, and null does not establish release availability.
Reported context-length value`result.data.code_leaderboard[].contextLength`	`number \| null`	Source integer or null; 20 saved coding entries are null. Units and supported usage conditions are not defined by this endpoint, so preserve the value without a capacity guarantee.
Reported input-price value`result.data.code_leaderboard[].inputPricePerMillion`	`number \| null`	Source number or null; eight saved coding entries are null. The field name says per million, but currency, priced quantity and billing conditions are not specified. Null is unknown, not free.
Reported output-price value`result.data.code_leaderboard[].outputPricePerMillion`	`number \| null`	Source number or null, separate from input price. Currency and billing conditions are unspecified, and this is not a verified current vendor quote. Retain nulls.
Separate Text Arena snapshot`result.data.text_leaderboard_snapshot`	`array`	Source-described top-ten text collection; ten entries are observed in the saved response. The same field names belong to text context and must not be substituted for coding records or merged into one ranking.
Text source model key`result.data.text_leaderboard_snapshot[].modelKey`	`string`	Original text-collection identifier. Keep it with this collection, even if a coding entry has a similar display name or the same source key.
Source text rank`result.data.text_leaderboard_snapshot[].rank`	`number`	Integer source position in the text snapshot, not a coding rank or a directly interchangeable task result.
Source text rating`result.data.text_leaderboard_snapshot[].rating`	`number`	Numerical rating for the separate text collection. A shared scale with coding ratings and source methodology are not established by this response.
Source text vote count`result.data.text_leaderboard_snapshot[].votes`	`number`	Source-reported integer count for the text snapshot. Its counting basis and independence are unspecified; do not add it to coding votes as a combined performance measure.

Additional notes

When choosing a coding model, start by comparing the records in code_leaderboard. This Tool also returns text_leaderboard_snapshot, described by the source as the top ten Text Arena models, for separate general-text context. No named input control is documented; request the combined data without inventing a model, year, filter or top-N parameter.

Keep every rank, rating and vote count attached to its collection and source modelKey. A shared display name does not establish an identical tested configuration, and a text rank is not a coding rank. Retain the source rankLower and rankUpper labels literally: their numbers are reversed relative to an ordinary ascending interval in most saved rows. Do not repair that convention or infer a confidence percentage.

Use the records to prepare candidates for your own evaluation. Preserve null metadata and blank organization labels; source license labels and model links are references rather than verified permissions or access. Context-length and price fields lack defined units or currency here, and no leaderboard-update time is supplied. Check appropriate current model evidence before drawing deployment, cost or project-suitability conclusions.

Questions about this tool.

Inputs, results and details to know before running it.

Get connected

It is the position reported by that source collection. The Tool reads leaderboard records; it does not run your tasks, measure repository-specific correctness or prove that one model is best for every project.

Review LMArena coding rankings with text context

Review LMArena coding rankings with text context

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Ask for the coding rankings

Read the two rankings separately

Choose candidates for your own task

Questions about this tool.

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Ask for the coding rankings

Read the two rankings separately

Choose candidates for your own task

Questions about this tool.

Does first place identify the best model for my codebase?

Can I merge the two collections by model name?

How should I read uncertain rankings or missing model details?