DeepSeek-V2.5

DeepSeek · Released Sep 6, 2024
DFO score #508 / 697
10.9 leaderboard score
Speed
– Not reported
Input Price
– Not priced
Output Price
– Not priced
Context
– Not reported

Analysis Summary

DeepSeek-V2.5 is an older model with a low measured intelligence result. The supplied record contains no coding or agent benchmarks, instruction-following evidence, context capacity, speed, or pricing. As a result, there is little data to assess its reliability on software tasks, structured content, or long documents, and no demonstrated operating advantage to balance that uncertainty.

It could be included in a controlled test for simple text tasks if a team can check outputs and compare the actual cost of running it. The information provided does not support adopting it as a default for client-facing content, complex analysis, or tool-driven workflows. Teams should favor alternatives with more recent and task-specific evaluation data when quality matters.

Assessed September 23, 2026

Editorial notes

DeepSeek-V2.5 has limited measured capability in the supplied record, with no task-specific coding, tool-use, context, or pricing evidence for production use.

The DFO score ranks models on the Artificial Analysis Intelligence Index and is refreshed as new models launch. The write-up and verdict are our own. Feedback?

DFO Verdict

DeepSeek-V2.5 has limited measured capability in the supplied record, with no task-specific coding, tool-use, context, or pricing evidence for production use.

#508 of 697 overall Down 1 this week

Benchmarks

Intelligence Index 6.6 #326 of 435 · Artificial Analysis
Coding Index Not published yet Artificial Analysis
Agentic Index Not published yet Artificial Analysis
Creative Writing Not rated Arena rating
Instruction Following Not rated Arena rating

Artificial Analysis data refreshed Oct 8, 2026. Arena ratings from the Oct 2, 2026 text leaderboard, with style control.

Artificial Analysis has not published coding or agentic results for DeepSeek-V2.5 yet. Until it does, the model sits below tested models on the Coding and AI Agents lists.

How DeepSeek-V2.5 compares

DeepSeek-V2.5 ranks #326 of 435 AI models we track for overall intelligence. DeepSeek-V2.5 is currently free to use via OpenRouter.

Position in the field
DFO score: ahead of 28% of models #502
worst in fieldmedianbest in field
Price vs frontier peers · $ per 1M tokens
DeepSeek-V2.5 $0.00 in $0.00 out
Anthropic: Claude Opus 5.5 $4.00 in $20.00 out
Anthropic: Claude Sonnet 5.5 $2.00 in $10.00 out
Anthropic: Claude Fable 5.1 $10.00 in $50.00 out

Dark bar = input · light bar = output, scaled to the priciest peer.

Context window vs peers · tokens

1M tokens ≈ 8 full-length novels or ~2,500 pages of business documents in a single request.

Frequently asked questions about DeepSeek-V2.5

How much does DeepSeek-V2.5 cost?

DeepSeek-V2.5 is currently available for free via OpenRouter.

Who created DeepSeek-V2.5?

DeepSeek-V2.5 is developed by DeepSeek and was released on September 6, 2024.

© 2026 Design for Online Ltd. Registered in England and Wales No. 10328553. VAT Registered. Design for Online® and Forerunner® are registered trademarks.