Back to Models
DeepSeekOpen Source

DeepSeek-V3.1 Terminus

DeepSeek-V3.1 Terminus, developed by DeepSeek-AI, is a large foundation model boasting approximately 6.7 trillion parameters. It features a context window of 128K and provides advanced chat capabilities.

Parameters

6710.0B

Context Window

128K

License

MIT

Release Date

2025-09-22

API Pricing

API pricing for this model is not yet available

Strengths

  • Overwhelming parameter scale
  • Long context window
  • Open MIT license

Weaknesses

  • Very large model size
  • Requires extensive storage capacity
  • High computational resource demand

Use Cases

  • Development of advanced conversational AI
  • Analysis of large-scale data
  • Understanding and processing complex context

Deep Analysis

Parameters

684.5B MoE

GPQA Diamond

74.9%

SWE-Bench Verified

66.0%

MMLU-Redux

91.8%

Pricing

$0.30/$0.95 per 1M tokens

Release Date

September 22, 2025

Strengths

  • Strong coding agent (66% SWE-Bench)
  • Excellent general knowledge (91.8% MMLU)
  • Extended long-context training
  • MIT license

Weaknesses

  • Non-thinking mode only
  • AIME 2025 modest (49.8%)
  • Trails frontier by 10+ pp
  • Quickly superseded

Competitor Comparison

ModelGPQAPrice
V3.180.1%$0.28/$0.42
V3.282.4%$0.28/$0.42

DeepSeek-V3.1 Terminus is an extended pre-training variant with 684.5B params trained on 6.7T tokens.

Analysis generated: 2026-05-24