deepseek

DeepSeek V2.5

Released 2024-09-05Context 128Kdeepseek-v2-5

PipelineScore

77.4MAINLINE

Ranked #30 of 52 models · 42nd percentileAn even spread: no standout, no liability (74 to 78 across all five categories). Best-fit profile: Agentic.

Category breakdown

Score per category, normalized 0–100 against the v1 anchor.

Code

77.3

Reason

78.3

Tool Use

78.0

RAG

77.5

Speed

74.3

Reason78.3

Tool Use78.0

RAG77.5

Every submission of DeepSeek V2.5 on the 0–100 scale. The spread is the point: where it runs changes what you get.

0255075100

Best 78.6 on lab-4x-rtx-3090 · lowest 76.6 on m2-ultra-192gb · spread 2.0 pts across 4 runs. Hover a dot for its rig.

A taste of what the test pack measures. Full prompts are private and rotated daily.

CodeDifficulty 1code-fib-1

Write a Python `fib(n)` returning the nth Fibonacci number, O(n).

ReasonDifficulty 1reason-math-1

Two trains, opposite directions, given speeds and start times — when do they meet?

RAGDifficulty 2rag-extract-1

From the context, extract net sales, operating margin, and free cash flow as a JSON object. Numbers only.

Tool UseDifficulty 2tool-schema-1

Given an OpenAPI schema with limit/offset/sort, fill JSON for 'next 50, recent first.'

RAGDifficulty 2rag-grounding-1

Context lacks the answer — does the model fabricate or correctly say it can't?