deepseek-v3.1-thinking

DeepSeekPlus

Use this model Try it

Availability100.0%

Available

Samples: 16

Latency973 ms

Output speed100.7 tok/s

Input / 1M tokens: $0.15
Output / 1M tokens: $0.75
Route: deepseek_officialPlus routes

Roleplay

Scores from deepseek-v3.1

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Creative Writing v3
deepseek-ai/DeepSeek-V3.1: 1,436.1
145 Models · 2026-10-25

6.6/ 10

Routes

Score

Ability score divided by token cost, ranked across measured model routes. Cost uses equal input and output token counts; higher is better value.

Sort by

						Use this model
deepseek_official	$0.15	$0.75	$0.075	6.6/10	8.9/10	Plans and tiers

Quick start

Access and billing

curl --fail-with-body 'https://api.tokenreply.com/v1/chat/completions' \
  -H "Authorization: Bearer $TOKENREPLY_API_KEY" \
  -H 'Content-Type: application/json' \
  --data '{
  "model": "deepseek-v3.1-thinking",
  "messages": [
    {
      "role": "user",
      "content": "Reply with hello."
    }
  ]
}'

Specifications

Context window: Not listed
Maximum output: Not listed
Knowledge cutoff: Not listed
Input formats: Not listed
Output formats: Not listed
API formats: openai

Benchmarks

Roleplay

Scores from deepseek-v3.1

Roleplay combines public evaluations of emotional understanding, creative writing and story-related skills. It is a comparative reference, not a direct roleplay test.

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Creative Writing v3
deepseek-ai/DeepSeek-V3.1: 1,436.1
145 Models · 2026-10-25

6.6/ 10

Reasoning

Scores from deepseek-v3.1

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Artificial Analysis Intelligence
deepseek/deepseek-chat-v3.1: 13.7
217 Models · 2026-10-25

2.7/ 10

Coding

Scores from deepseek-v3.1

Each benchmark is rescaled to 1–10 within its own leaderboard. We average results from the same publisher first, then give each publisher equal weight. Only verified model matches are included.

Design Arena · Code
deepseek/deepseek-chat-v3.1: 1,110
118 Models · 2026-10-25

5.8/ 10

TokenReply

deepseek-v3.1-thinking

Routes

Quick start

More DeepSeek models