Skip to main content
Glama

compare evals

compare_evals
Read-onlyIdempotent

Compare two finished eval runs arm-by-arm: enter a baseline run and an after run to get per-arm delta, 95% interval, and a significance flag instead of reading two reports side by side.

Instructions

Compare two finished eval runs arm-by-arm (before vs after a prompt, tool or index change) and get each arm's delta with a 95% interval and a significance flag, instead of eyeballing two reports. GET /v1/evals/compare (API-key scope: read). Returns: {a:{id, name, created_at, judge_model, eval_kind}, b:{…same}, rows:[{arm (candidate key), metric ("win_rate" for comparison runs; "observed_pass_rate" and, when both runs carry one, "corrected_pass_rate" for criterion runs), a:{rate, n, ci:[lo,hi]}, b:{rate, n, ci}, delta (b.rate − a.rate), delta_ci:[lo,hi] (Newcombe 95%), significant (interval excludes zero)}], unmatched:{a:[arm keys only in a], b:[arm keys only in b]}}. Cache-Control: no-store. Notes: 400 when a or b is missing. 404 when either run is not in this workspace. 412 {code:"precondition_failed"} when both runs are not DONE, when they are different eval kinds, or when they used different judge models (a delta between judges measures the judges, not your change — re-run one with the other's judge). Best used with identical sample_filters on both runs. Read-only, no spend.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
aYesEval run id of the BEFORE run (baseline of the comparison).
bYesEval run id of the AFTER run. Delta is reported as b − a.
Install Server

TDQS

A4.8/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already mark readOnly/idempotent/non-destructive, and the description adds substantial behavioral detail: exact response shape, Cache-Control: no-store, sign convention (delta = b − a), Newcombe 95% interval, significance rule, no spend, and precondition failure semantics. No contradiction with annotations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is long but densely informative and well structured: purpose, endpoint, response shape, error conditions, best-use guidance, and safety note. Every clause earns its place, and the core purpose is front-loaded.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema present, the description fully specifies the return object including nested fields, arm keys, metrics, confidence intervals, unmatched sets, and error cases. It also covers operational context (workspace scoping, judge-model precondition, sample_filters), making it complete for correct invocation.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Although schema coverage is 100%, the description adds the crucial directional meaning: a is the BEFORE/baseline run and b is the AFTER run, with delta explicitly defined as b.rate − a.rate. This goes beyond the schema's short descriptions and prevents a common misuse.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description states a specific verb and resource ('Compare two finished eval runs') and distinguishes the tool from single-run eval tools by explaining the arm-by-arm delta output. The phrase 'instead of eyeballing two reports' makes the intent unambiguous.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It clearly states when the tool applies (finished runs, same eval kind, same judge model, ideally identical sample_filters) and documents rejection conditions (400/404/412), including the important 'delta between judges measures the judges' caveat. It does not explicitly name sibling tools like get_eval for single-run needs, but the comparison context is otherwise well specified.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Other Tools

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/omnia-v/errorbar-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server