DeepSeek V3.1 Terminus Model Specs, Costs & Benchmarks (July 2026)
DeepSeek V3.1 Terminus, developed by DeepSeek, features a context window of 163.8K tokens. The model costs $0.27 per million tokens for input and $1.00 per million tokens for output. It was released on September 22, 2025, and has achieved impressive scores in various benchmarks.
Access DeepSeek V3.1 Terminus & 210+ other AI models all in one platformTry Magica for free
Overview
 DeepSeek V3.1 Terminus | |
|---|---|
Model Provider The organization behind this AI's development | DeepSeek |
Input Context Window Maximum input tokens this model can process at once | 163.8K tokens |
Output Token Limit Maximum output tokens this model can generate at once | 32.8K tokens |
Release Date When this model first became publicly available | September 22, 2025 10 months ago September 22nd, 2025 |
Knowledge Cutoff Latest training-data date reported by the provider | March 31, 2025 |
Pricing
| DeepSeek V3.1 Terminus | |
|---|---|
Input Token Cost Cost per million input tokens | $0.27 per million tokens |
Output Token Cost Cost per million output tokens | $1.00 per million tokens |
Cache Read Cost Cost to reuse cached input tokens | $0.14 per million tokens |
Capabilities & Features
| DeepSeek V3.1 Terminus | |
|---|---|
Input Types Supported input formats | text |
Output Types Supported output formats | text |
Tokenizer Text encoding system | DeepSeek |
Key Features Advanced capabilities | ✓ Function Calling✓ Structured Output✓ Reasoning Mode |
Reasoning Controls Defaults and configurable reasoning effort | Optional |
Open Source Model availability |
Benchmarks
| DeepSeek V3.1 Terminus | |
|---|---|
Intelligence Index Overall model quality across independent evaluations | 30.4 (Artificial Analysis index; higher is better) |
Coding Index Programming performance across independent evaluations | 43.5 (Artificial Analysis index; higher is better) |
Agentic Index Ability to complete multi-step agentic tasks | 18.1 (Artificial Analysis index; higher is better) |
Best Design Arena Score Highest human-preference Elo score across design arenas | 1,220 Elo (#41 in UI Component (59.3% win rate)) |
Compare This Model
See how DeepSeek V3.1 Terminus compares with other top models
Compare DeepSeek V3.1 Terminus with top models in each category:
🚀Programming
Best models for coding and development
🎨Creative & Roleplay
Models optimized for creative writing
📢Marketing
Content creation and marketing tasks
💻Technology
Technical analysis and explanations
🔬Science
Scientific research and analysis
🌐Translation
Multilingual translation tasks
