Skip to main content
AITier

DeepSeek-V2.5 (Dec '24)

DeepSeekDeepSeekOpen Weightdeepseek

Description

DeepSeek-V2.5 is an upgraded version that combines DeepSeek-V2-Chat and DeepSeek-Coder-V2-Instruct, integrating general and coding abilities. It better aligns with human preferences and has been optimized in various aspects, including writing and instruction following.

Release Date
2024-12-10
Parameters
236.0B
Context Length
164K
Modalities
text

Capability Radar

28
general
60
coding
63
reasoning
42
science
62
agents
0
multimodal

Rankings

Domain#RankScoreSource
General Ranking487
29.0
AA
Science394
38.0
AA

Benchmark Scores (LLM Stats)

(LLM Stats (zeroeval))

Chat

MT-Bench0.90 / 100SR

General

MMLU80.4%SR
AlignBench80.4%SR
DS-FIM-Eval78.3%SR
Arena Hard76.2%SR
AlpacaEval 2.050.5%SR

Math

GSM8k95.1%SR
MATH74.7%SR

Reasoning

HumanEvalOpenAI (2021)89.0%SR
BBH84.3%SR
HumanEval-Mul73.8%SR
Aider72.2%SR
DS-Arena-Code63.1%SR
LiveCodeBench(01-09)41.8%SR
SWE-Bench Verified16.8%SR

AA Evaluation Indices

(Artificial Analysis)
Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))
76.3
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))
66.6
Gpqa(NYU + Cohere + Anthropic (2023))
42.3
Intelligence Index(Artificial Analysis)
6.6

LLM Stats Category Scores

(LLM Stats (zeroeval))
Roleplay
90
Communication
90
Chat
80
Language
80
Legal
80
Math
80
Finance
80
Healthcare
80
Reasoning
70
General
70
Creativity
70
Writing
70
Code
60
Frontend Development
20

Pricing

Input PriceFree
Output PriceFree
Blended Price (3:1)Free

Speed

Tokens/sec0.0
Time to First Token0.00s
Time to Answer0.00s

Provider Price Ranking

No provider data available

External Sources

LLM StatsArtificial Analysis