deepseek
DeepSeek: DeepSeek V3.1 Terminus
deepseek/deepseek-v3.1-terminus
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
deepseek/deepseek-v3.1-terminus is a text-to-text update of DeepSeek V3.1 that, according to its model documentation, preserves the prior model's capabilities while improving language consistency and its Code Agent and Search Agent. OpenRouter lists a 163,840-token context, tool calling, structured outputs, and benchmark/catalog data. Vendor-reported results improve on several agentic tasks but regress on some coding and Chinese browsing measures, so the evidence supports cautious evaluation rather than a blanket upgrade claim.[1][2]
Strengths
- Maintains V3.1 capabilities while targeting user-reported mixed Chinese-English text, occasional abnormal characters, and agent behavior.[1][2]
- Catalog metadata indicates a 163,840-token context, text input/output, and parameters for tools, tool choice, response format, and structured outputs.[1]
- Supplied benchmark tables record gains over V3.1 on several agentic evaluations; OpenRouter also lists design-arena results and coding and agentic indexes.[1][2]
Limitations
- OpenRouter lists text input and output only; the source pack combines catalog metadata with vendor documentation and supplies no independent workload validation.[1][2]
- Vendor-reported comparisons are not uniformly positive: results regress against V3.1 on Codeforces, Aider-Polyglot, and BrowseComp-zh.[2]
- The current checkpoint has a known issue in which self_attn.o_proj parameters do not conform to the UE8M0 FP8 scale format.[2]
Where to use it
- Consider for long-document drafting, summarization, or QA when a large text-only context is useful; verify effective long-context behavior.[1]
- A plausible fit for tool-using coding assistants and terminal-oriented automation; test repository-level reliability before rollout.[1][2]
- Worth piloting for search-assisted research agents, while separately checking English and Chinese browsing because reported results moved in opposite directions.[2]
Sources
- [1] OpenRouter catalog · DeepSeek: DeepSeek V3.1 TerminusSources checked: Nov 13, 2026
- [2] Model documentation · DeepSeek: DeepSeek V3.1 TerminusSources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
DeepSeek-V3.1 Terminus is an update to [DeepSeek V3.1](/deepseek/deepseek-chat-v3.1) that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further optimizing the model's...
Specifications
- Model ID
- deepseek/deepseek-v3.1-terminus
- Creator
- deepseek
- Added to OpenRouter
- 22.09.2025
- Context length
- 163,840
- Modality
- text->text
- Input modalities
- text
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.270 / out $1
- Knowledge cutoff
- 31.03.2025
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | No measurement | reasoning: optional · parameters: include_reasoning, reasoning |
| Coding | BB — next 30% of the cohort | — |
| Agents | CC — below the B band | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | reasoning: optional · parameters: include_reasoning, reasoning |
| Document analysis | No measurement | — |
| Vision | No measurement | — |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Artificial Analysis | Coding index | 43.5 | BB — next 30% of the cohort | 13.11.2026 |
| Artificial Analysis | Agentic index | 6.8 | CC — below the B band | 13.11.2026 |
| Design Arena | models · 3dwin rate 56% · rank 81 | 1,123Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · codecategorieswin rate 56% · rank 76 | 1,166Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · datavizwin rate 53.4% · rank 74 | 1,166Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · gamedevwin rate 52.5% · rank 92 | 1,121Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · svgwin rate 50.1% · rank 79 | 1,057Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · uicomponentwin rate 59.5% · rank 71 | 1,173Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · websitewin rate 56.4% · rank 76 | 1,180Elo | BB — next 30% of the cohort | 13.11.2026 |
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| SiliconFlowfp8 | 163,840 | $0.270 | $1 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.