Skip to main content
DeepSeek

DeepSeek V3.1 Terminus

DeepSeek V3.1 Terminus is a chat model from DeepSeek, released as an updated point revision in the company's DeepSeek V3.1 line.

Input from
$0.270 / 1M tokens
across 2 providers

API Pricing

ProviderInput / 1MOutput / 1MCached / 1M
$0.270$1.00$0.135
$0.270$1.00$0.135

Prices updated daily. Last check: Sep 2, 2026

DeepSeek V3.1 Terminus pricing by provider

Input, output, and batch rates, plus alternatives, for one provider at a time.

Performance & Benchmarks

Source: Artificial Analysis →
Intelligence
21.7 / 100
Math
53.7 / 100

Reasoning & Knowledge

  • MMLU-Pro83.6%
  • GPQA Diamond75.1%
  • Humanity's Last Exam8.7%

Coding

  • LiveCodeBench52.9%
  • SciCode32.1%

Math

  • AIME 202553.7%

Agentic & Tool Use

  • Terminal-Bench Hard31.8%
  • τ²-bench37.1%

Instruction & Long Context

  • IFBench41.2%
  • Long-Context Reasoning45.0%

Benchmarks measured Sep 2026. Scores are independent evaluations, not vendor-reported.

Model Details

General

Creator
DeepSeek
Modalities
Text

Capabilities

Tool Calling
No
Open Source
No

Strengths & Limitations

Strengths

  • Point revision of DeepSeek V3.1, so prompts and integrations built for V3.1 generally carry over with minimal changes
  • Available through several inference providers, allowing side-by-side comparison of throughput and cost in the pricing table on this page
  • Standard chat-completions interface, which makes it a drop-in candidate for existing OpenAI-compatible client code used by many providers
  • Part of the widely deployed DeepSeek V3 series, so community documentation, prompt patterns, and integration examples are readily available
  • Multi-provider availability reduces single-vendor dependency for teams that need failover or regional routing
  • Positioned as a general-purpose chat model rather than a narrow specialist, covering conversational, drafting, and coding-assistance workloads

Limitations

  • We do not track a confirmed context window for this revision — check the serving provider's documentation before designing long-context workloads
  • No verified benchmark scores in our records for DeepSeek V3.1 Terminus, so capability comparisons must rely on third-party evaluations
  • Throughput and time-to-first-token figures are unpopulated in our benchmark data, so speed cannot be compared from this page alone
  • As a point revision, the differences from DeepSeek V3.1 are incremental and may not justify a migration on their own
  • Behavior, context limits, and feature support (such as tool calling) can vary between the providers that host it

Key Features

Chat-completion API interface
DeepSeek V3.1 series lineage as a revision-level update
Text generation for conversational, drafting, and coding-assistance tasks
Availability across multiple third-party inference providers
Compatibility with tooling written for earlier DeepSeek V3.1 endpoints
Provider-dependent context window and rate limits
Cross-provider price and availability comparison via the pricing table on this page

About DeepSeek V3.1 Terminus

DeepSeek V3.1 Terminus is a large language model developed by DeepSeek and offered through inference APIs as a chat-completion model. It sits within the DeepSeek V3 series as a maintenance-style update to DeepSeek V3.1 — the "Terminus" suffix marks a revision of the same base model rather than a new generation, so it is positioned as a refinement of V3.1 behavior rather than a separate architecture. Our database records DeepSeek V3.1 Terminus as a text chat model; we do not currently track a confirmed context window length, modality list, or published benchmark scores for this specific revision, and the throughput and latency figures in our records are unpopulated rather than measured at zero. Because DeepSeek V3.1 Terminus is served by multiple third-party inference providers in addition to DeepSeek's own API, practical characteristics such as maximum context length, output token limits, tool-calling support, and tokens-per-second throughput can differ between providers, and the provider's own documentation is the authoritative source for those details. In practice, models in the DeepSeek V3.1 family are used for general assistant work, coding help, and long-form text generation, and they are frequently evaluated by teams looking for an alternative to the closed North American chat model families. Buyers comparing DeepSeek V3.1 Terminus against DeepSeek V3.1 itself should treat them as close relatives — the choice usually comes down to which revision a given provider hosts, and at what throughput and rate limits.

Common Use Cases

DeepSeek V3.1 Terminus suits general-purpose chat and text workloads: assistant and support interfaces, drafting and rewriting documents, summarization, code explanation and generation, and data transformation where output is returned as text. Because it is a revision of DeepSeek V3.1, it is a natural candidate for teams already running V3.1 who want to test whether the updated checkpoint behaves better on their own evaluation set without rewriting prompts. It is also commonly evaluated as a cost-and-throughput alternative to closed chat models on high-volume text pipelines. For workloads that depend on very long inputs, structured tool calling, or image input, confirm support with the specific provider you plan to use, since we do not track those fields for this revision.

Frequently Asked Questions

How much does DeepSeek V3.1 Terminus cost?

Pricing depends on which provider serves the model and on the pricing type — per-input-token, per-output-token, batch, or dedicated capacity all differ. Rates also change frequently. See the pricing table on this page for the current figures across the providers we track.

What is DeepSeek V3.1 Terminus best used for?

It is a general-purpose chat model, so it fits assistant and chatbot interfaces, drafting and summarization, coding assistance, and high-volume text processing. It is a particularly convenient choice for teams already running DeepSeek V3.1 who want to compare an updated checkpoint on their existing prompts.

How is DeepSeek V3.1 Terminus different from DeepSeek V3.1?

The "Terminus" label marks a point revision of the same V3.1 line rather than a new model generation. Expect incremental changes in behavior rather than a different architecture or a different integration path, and validate on your own evaluation set to see whether the revision helps your workload.

What context window does DeepSeek V3.1 Terminus support?

We do not have a confirmed context window recorded for this revision, and the practical limit often depends on the serving provider. Check the documentation of the provider you select before building workloads that depend on very long inputs.

Does DeepSeek V3.1 Terminus support tool calling or image input?

Our metadata records it as a text chat model and does not confirm tool calling or vision support for this revision. Feature availability can also vary by provider, so verify with the specific API endpoint you intend to use.