Skip to content

DeepSeek

13 models · avg score 64

Last updated: 28m ago
Models13
Avg Score64
Best Rank#60

Provider Strengths

Avg Composite Score
64
Best Model
DeepSeek V4 Pro
Score 87
Price Range (per 1M out)
$0.18 - $2.50
Avg Context Window
424K
Total Models
13
Capabilities Coverage
4/7
capabilities present

Model Timeline

Aug 2026DeepSeek V4 Flash LatestScore 40
Jul 2026DeepSeek V4 Flash 0731Score 40
Apr 2026DeepSeek V4 ProScore 87
Apr 2026DeepSeek V4 Flash 0423Score 40
Dec 2025DeepSeek V3.2Score 81
Sep 2025DeepSeek V3.2 ExpScore 72
Sep 2025DeepSeek V3.1 TerminusScore 69
Aug 2025DeepSeek V3.1Score 72
May 2025R1 0528Score 79
Mar 2025DeepSeek V3 0324Score 72
Jan 2025R1 Distill Llama 70BScore 41
Jan 2025R1Score 74
Dec 2024DeepSeek V3Score 70

Pricing Distribution

Free
0models
Budget
11models
Standard
2models
Premium
0models

Capability Coverage

Vision
0/13
Function Calling
12/13
Reasoning
11/13
JSON Mode
12/13
Streaming
13/13
Web Search
0/13
Image Output
0/13

Recently Added (2)

Last 30 days

Benchmark Leaders

Arena Elo
DeepSeek V4 Pro
1463 elo
MMLU
R1 0528
91.5%
MMLU-Pro
DeepSeek V4 Pro
87.5%
GPQA Diamond
DeepSeek V4 Pro
90.1%
HumanEval
DeepSeek V3 0324
84.5%
MATH-500
R1 0528
97.8%
SWE-bench
DeepSeek V4 Pro
80.6%

Verification & Source Coverage

Put catalog freshness, service status, and official links in one place so this provider page reads like a monitored profile instead of a static directory.

Catalog verified
Aug 8, 2026

Backed by the live model catalog cache.

Status verified
Aug 8, 2026

Backed by the provider status feed.

Latest launch
Aug 1, 2026

Tracking back to Dec 26, 2024

Source mix
13 live

No embedded product models

Operational Health

Track reliability alongside pricing and model depth so provider selection stays grounded in operating reality.

Full status page
Current status
operational

All Systems Operational

30d uptime
100%

Rolling availability across the last 30 days.

Stability score
100

Composite reliability indicator from recent status behavior.

Incidents in 30d
0

Includes degraded performance, partial outages, and major incidents.

Provider Catalog

Sort:
#ModelScore
60DeepSeek V4 Pro
87
81DeepSeek V3.2
81
90R1 0528
79
134R1
74
150DeepSeek V3.2 Exp
72
151DeepSeek V3.1
72
152DeepSeek V3 0324
72
163DeepSeek V3
70
166DeepSeek V3.1 Terminus
69
232R1 Distill Llama 70B
41
235DeepSeek V4 Flash Latest
40
236DeepSeek V4 Flash 0731
40
275DeepSeek V4 Flash 0423
40
13 models

Scoreboard

RankModelScore
#60DeepSeek V4 Pro87
#81DeepSeek V3.281
#90R1 052879
#134R174
#150DeepSeek V3.2 Exp72
#151DeepSeek V3.172
#152DeepSeek V3 032472
#163DeepSeek V370
#166DeepSeek V3.1 Terminus69
#232R1 Distill Llama 70B41
#235DeepSeek V4 Flash Latest40
#236DeepSeek V4 Flash 073140
#275DeepSeek V4 Flash 042340
Frequently Asked Questions

DeepSeek currently offers 13 AI models on our platform, with an average quality score of 64/100.

The highest-rated DeepSeek model is DeepSeek V4 Pro, ranked #60 overall with a composite score of 87/100.

DeepSeek models have an average score of 64/100. Their best model, DeepSeek V4 Pro, is ranked #60 globally. Visit our provider comparison page for detailed analysis.

DeepSeek Models - Rankings, Pricing & Benchmarks (2026) | LM Market Cap