Compare models on the LMArena search leaderboard

Read Search Arena model rankings without a required query. Compare source positions, ratings and vote counts to choose models for your own search tests.

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-search-leaderboard-lmarena to read Search Arena rankings and compare the model or organization I name. 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. The ranking read needs no extra input; handle the request internally without inventing search-query, model, provider, category, count or pagination filters. Select my requested model or organization only from actual returned source records after retrieval; use source identity and organization context, not a fabricated record or guessed provider classification. If absent or ambiguous, report absence or show readable choices. Apply an ordinary comparison count to returned records only. Show model names, source ranks, scores, vote counts and actual reference links, retaining real source identities internally and distinguishing variants. Preserve original ranking-range labels and values, including reversed numerical order, without assigning confidence; keep any additional source ranking value separately labelled with its calculation unspecified. Leave missing model details and prices unknown, and verify currency and pricing conditions separately. This reads ranking records, not webpages or search answers, and does not measure my query’s factual accuracy, citation quality or retrieval completeness. Bracketed example text is replaced by my own content.

Task:
Read the LMArena search-model leaderboard. Compare [model or organization to compare, such as Perplexity] with the first [number of models to compare, such as three] models returned. Show names, ranks, scores and votes; explain limits for search accuracy and citations.

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

Compare models on the LMArena search leaderboard

Read Search Arena model rankings without a required query. Compare source positions, ratings and vote counts to choose models for your own search tests.

You receive

Search Arena model records with readable names, organizations, reference links, source ranks, scores and vote counts. Use the returned records to choose models for separate search tests, keeping missing information clear.

Start with your task

Read the LMArena search-model leaderboard. Compare [model or organization to compare, such as Perplexity] with the first [number of models to compare, such as three] models returned. Show names, ranks, scores and votes; explain limits for search accuracy and citations.

What to know

This retrieves a leaderboard; it does not perform a web search or test a listed model. The records alone do not establish factual accuracy, citation quality or retrieval completeness for your query. Source update time and evaluation details are absent from the saved response. Keep rank bounds as supplied, including reversed numeric order; verify model pricing, limits and license terms separately.

Inputs & outputs

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

Input schema

Input schema
Field / Path	Type	Requirement	Meaning
Optional query settings`queryParams`	`object`	Optional	Optional open object with no documented named choices or defaults. No search phrase is required; named search, model, category, pagination and top-N controls are not defined.

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
Search Arena model records`result.data.entries`	`array`	Source-ordered records for the search/retrieval leaderboard. The saved response has 34 records; neither a fixed size nor a complete model catalogue is established.
Source model key`result.data.entries[].modelKey`	`string`	Identity supplied for matching a model across fields. Keep the exact key; a display name or organization is not a substitute.
Model display name`result.data.entries[].modelDisplayName`	`string`	Name used in the source record. Read it alongside modelKey and the reference link; it is not proof of a current available version.
Reported organization`result.data.entries[].modelOrganization`	`string`	Organization text attached to this record. It is source attribution rather than a verified ownership, endorsement or service-access statement.
Model reference link`result.data.entries[].modelUrl`	`string`	Source-supplied model link for separate checking. Its presence does not verify the linked page, current pricing or permissions.
Source position`result.data.entries[].rank`	`number`	Position reported on this search leaderboard. Compare within the returned collection; it is not a score for your own query.
Reported rating`result.data.entries[].rating`	`number`	Numeric source rating, accompanied by bounds and vote count. Scale and evaluation population are unspecified; do not convert it to an accuracy rate.
Reported vote count`result.data.entries[].votes`	`number`	Integer vote count supplied for this record. The response does not identify unique voters, vote dates or the tested query population.
rankLower source bound`result.data.entries[].rankLower`	`number`	Preserve the named source value. It exceeds rankUpper numerically in all 34 saved records; no interval interpretation or relabelling is justified.
rankUpper source bound`result.data.entries[].rankUpper`	`number`	Read with rankLower under its original name. Do not swap the two values to create an assumed lower-to-upper interval.
Lower reported rating bound`result.data.entries[].ratingLower`	`number`	Source value paired with ratingUpper. Keep the reported number without inventing a confidence percentage or statistical method.
Upper reported rating bound`result.data.entries[].ratingUpper`	`number`	Other rating-bound value provided for the same model. Its presence alone does not establish significance against another model.
rankStyleControl source value`result.data.entries[].rankStyleControl`	`number`	Integer supplied in all 34 saved records. No calculation, style adjustment or evaluation method is documented; keep it separate from rank.
Reported context value`result.data.entries[].contextLength`	`number \| null`	Numeric in 28 saved records and null in six. Unit and effective input/output limits are not supplied here; verify them before planning a workflow.
Reported input-price value`result.data.entries[].inputPricePerMillion`	`number \| null`	Source price-per-million label: 30 saved records contain numbers and four are null. Currency, priced quantity and conditions are unspecified; this is not the Toolmu call fee.
Reported output-price value`result.data.entries[].outputPricePerMillion`	`number \| null`	Source price-per-million label with 30 numeric and four null saved values. Confirm the relevant model billing terms rather than deriving a payable quote.
Source license label`result.data.entries[].license`	`string`	Reported license text. It neither reproduces the license terms nor confirms that your intended use or service access is permitted.

Additional notes

This endpoint reads the Search Arena category for search and retrieval tasks. No search phrase, model selection or page size is required. The optional query container defines no named controls or defaults, so do not invent a query parameter from the task you want a model to solve. Choose which returned records to show after retrieval. For a Perplexity comparison, select its returned records alongside other models after retrieval; this is not a provider-only leaderboard.

Keep source rank, rating and votes attached to the same model key. The saved collection contains 34 records; that count is one observation, not a fixed response size or proof of complete model coverage. All saved rankLower values exceed rankUpper numerically. Preserve their original names and values rather than repairing or relabelling the source interval. The calculation and use of rankStyleControl are not supplied.

Context-length and model-price values are auxiliary source information. The saved result reports context values for 28 models and input/output prices for 30, with the rest null. It does not define context units, price currency, billing conditions, source-update time or evaluation population. Check model documentation and test representative queries before choosing a service; a leaderboard result is not a retrieved answer or a verified citation.

Questions about this tool.

Inputs, results and details to know before running it.

Get connected

The source supplies an additional value without explaining its calculation or evaluation method. Keep it separate from the main rank; it is not an accuracy percentage or answer-quality guarantee. Evaluate models with your own search tasks.

Compare models on the LMArena search leaderboard

Compare models on the LMArena search leaderboard

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Ask for the search-model rankings

Choose models to compare by name

Test your own search needs separately

Questions about this tool.

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Ask for the search-model rankings

Choose models to compare by name

Test your own search needs separately

Questions about this tool.

Can an extra ranking value prove which search model is better?

Is the retrieval time the leaderboard update time?

Are the reported model prices a current service quote?