Back to leaderboard
Compare companies
Side-by-side MCP results for companies evaluated by Claude Code. Compare score, rank, discoverability, cost, errors, runtime, and the underlying eval artifacts.
Model
Claude Code
CodexChoose 2-4 companies
Selected
Selected
Deepgram
Voice STT
Prefect
Durable Workflow
Summary
Compare status
Comparable
Comparable
Overall score
30
43
BestGrade
D
D
Category
Voice STT
Durable Workflow
Overall rank
#51
#42
BestCategory rank
#1
Best#1
BestFull eval
MCP Task Scoring
Task score
4%
24%
BestTask
Failed+0The agent attempted to use MCP tools but failed to complete the core task. Only 1 MCP tool call was made (ListAgents), w...
BestFailed+0The agent correctly used MCP tools exclusively for product operations and successfully completed step 1 (discovery). How...
BestUsed MCP Correctly
Failed+0MCP usage issues: Only 1 MCP tool call made (ListAgents), which is unrelated to transcription, No transcription MCP tool...
Passed+20Agent made 11 MCP tool calls
BestPartial Credit
Partial+1.5Partial credit: 15%
Partial+2.5Partial credit: 25%
BestZero Errors
Failed+0Failed tool calls: 13
BestFailed+0Failed tool calls: 10
BestEfficiency
Failed+026 tool calls (>25, inefficient)
BestFailed+044 tool calls (>25, inefficient)
BestTime
Partial+1.0414s (5-10min, very slow)
BestPartial+1.0509s (5-10min, very slow)
BestCost
Partial+2.0$0.45 ($0.40-0.70, expensive)
BestPartial+1.0$0.90 ($0.70-1.00, very expensive)
Runtime (actual)
6m 53s
Best8m 29s
Cost (actual)
$0.4500
Best$0.9015
Tool calls
26
Best44
Errors
13
10
BestTokens used
925,119
Best2,022,284
MCP Discovery
Discovery score
92%
Best89%
Official MCP
Passed5/5
BestPassed5/5
BestAuth Method
Passed15/15env var
BestPassed15/15env var
BestTool Count
Passed10/101 tools, lean
BestPassed10/1015 tools, lean
BestTool Description Length
Passed10/10247 tokens, lean
BestPartial7/104,185 tokens, moderate
llms.txt
Passed10/10
BestPassed10/10
BestOutput Schema
Failed0/5
BestFailed0/5
BestTool Stability
Passed5/5
BestPassed5/5
BestSecurity Posture
Partial9/10No API keys, tokens, or credentials are exposed in the tool definition or schema. The query parameter is a string type w...
BestPartial9/10No API keys or tokens in descriptions; workspace_id is properly typed as optional UUID; minor concern that workspace_id...
BestTool API Mapping
Partial14/15This tool maps 1:1 to a single semantic search API operation with a minimal, focused input schema (one query parameter)....
BestPartial14/15Tools map cleanly to individual API read operations with minimal bundling; each tool performs one primary action (get, s...
BestDescription Clarity
Partial14/15The description clearly explains the tool's purpose (semantic retrieval over Deepgram knowledge sources), what it return...
BestPartial14/15Descriptions are comprehensive and clear, explaining purpose, return values, filter operators, and providing multiple co...
BestEval Details
Integration task
# Task
Using the Deepgram MCP, transcribe audio.
1. Discover the MCP tools relevant to transcription
2. Transcribe this sample audio file:
https://storage.googleapis.com/aai-w...
# Task
Using the Prefect MCP, interact with the workflow engine:
1. Discover available MCP tools and their capabilities
2. List existing workflow executions (or deployments/flows...
Generated files
1
1
Trace tool calls
26
44
Completed
9/27/2026
10/4/2026
Stderr
None captured
None captured