readable AI benchmarks (simplified)

readable AI benchmarks

DeepSeek

DeepSeek V3 (Dec '24)

non-reasoning · open weights · Dec 26, 2024

DeepSeek V3 (Dec '24) has a Quality Score of 21.1, ranking 360th among 551 scored models. Per 1M tokens, pricing is $0.320 input and $0.890 output. It accepts text input, outputs text, and has a 128k token context window. Its Reliability Score is 29.2, ranking 313th among 553 scored models, below average (total average: 35.6).

Quality Score21.1coverage 100.0%: AA Intelligence and AA-Omniscience outcomes available
Value Scorenot available
Factual reliability29.2
Cache Discountnot available

Model specification

Reasoning
non-reasoning
Input modalities
text
Output modalities
text
Context window
128k tokens
Weights
open weights

Token prices USD per 1M tokens

Input
$0.320
Output
$0.890
Cached Input
not available
Cache write
not available
Cost per task
$0.020

Capability

Intelligence
8.5
Coding
not available
Agentic
not available
Omniscience
-41.6
Correct
25.4%
Blended price
not available

Answer outcomes

Fully graded outcomes (Correct + Incorrect): 92.5%

Correct 25.4%Incorrect 67.1%Partial / not attempted 7.5%

Artificial Analysis benchmarks

GDPval-AA v2
0%
τ³-Banking
5%
SciCode
36%
Humanity’s Last Exam
3%
GPQA Diamond
56%
CritPt
0%
AA-Omniscience
29%
AA-LCR
29%

Similar models

Support me! Patreon