DeepSeek V3.1 Terminus

Current board #41Arena #119Open weights

DeepSeek·deepseek/deepseek-v3.1-terminus

DeepSeek V3.1 Terminus is a language model published by DeepSeek. 5 OpenRouter endpoints are published in the latest snapshot. It accepts text input and returns text. The published context limit is 164K tokens. The catalogue declares support for reasoning, tool use, structured outputs.

ModelCap Index · Current board

64.5#41

Measured · 51% support · score interval 64.5–76.5

Market Gravity

33.6

Secondary market signal

Input

$0.27

per 1M tokens

Output

$1.00

per 1M tokens

Context

164K

tokens

Providers

endpoints

Gateway spend

—

latest · Vercel share

Providers

Uptime measured over the last 30 minutes

Provider	Input	Output	Context	Quant	Uptime
DeepInfrafp4Cheapest	$0.27	$0.95	164K	fp4	100.0%
AtlasCloudfp8	$0.30	$0.95	131K	fp8	100.0%
SiliconFlowfp8	$0.27	$1.00	164K	fp8	99.57%
Novitafp8	$0.27	$1.00	131K	fp8	100.0%
StreamLake	$0.343	$1.03	128K	—	99.49%

Hugging Face Inference Providers

Exact repository identity

Provider economics and telemetry published by the Hugging Face Router. These are operational signals, not benchmark results.

Provider	Input	Output	Context	TTFT	Throughput	Features
deepinfra live	$0.27	$0.95	164K	641 ms	56.6 tok/s	structured
novita live	$0.27	$1.00	131K	1462 ms	30.5 tok/s	tools · structured
featherless-ai live	—	—	—	—	—	—

Reasoning efforts

Best supported result can inform ModelCap

Published Max, X-High, High and other configurations for this model. These are sourced scores, not head-to-head essays.

Related community configurations

Human preference rating

Configuration	Rating	Source rank	Votes	Source
Thinking `deepseek-v3.1-terminus-thinking`	1417 CI 1407–1427	#112	3K	Arena Match 100%

Reported evaluation leads (1)

Excluded from score

Structured rows reported in this exact model repository at the pinned revision below. They are useful research leads, but remain quarantined until dataset revision, harness, configuration and identity pass ModelCap’s independent admission review.

Dataset / task	Metric	Value	Revision	Provenance
Idavidrein/gpqa diamond	—	74.2424	Unpinned `not published`	Reported Quarantined

ModelCap Index · Current board

Methodology

64.5

Measured · 2 public benchmark observations across 2 boards

Current board

#41

Public rank basis: Measured (measured)
Index support: 50.5% · moderate
Index interval: 64.5–76.5
Rank posterior: Not available
Top 5 / Top 10 probability: Not available / Not available
Posterior as of: Snapshot timestamp unavailable
Identity binding: Exact catalogue product · deepseek/deepseek/deepseek-v3.1-terminus/fp8 · aggregates disclosed benchmark configurations · endpoint configuration not claimed · artifact metadata deepseek-ai/DeepSeek-V3.1-Terminus@19510d6dc61f79dbd925bd51ee8a9081c509a4b6 (not the evaluation revision)
Direct benchmark observations: 2
Breadth-adjusted BES point estimate: 70.5

Bayesian Evidence Score · measured evidence input only

BES normalizes admitted public benchmark evidence for measured Index rows. Its legacy score and observables prior do not define the public language rank.

Observed capability: 70.5
Evidence status: confirmed · 73% mass
General preference: 70.1
Coding: 72.0
Agents & tools: —
Reasoning: —
Evidence breadth: 92%
Evidence coverage: 51%

Market and catalogue signals · secondary, never the language Index rank

Usage (OpenRouter popularity): 22.9
Liquidity (providers × uptime): 70.5
Open reach (HF downloads): 12.5
Surface (context / tools / modalities): 71.8
Freshness: 29.6

· OpenRouter popularity #81
· 5 live providers
· HF 30d downloads 17279
· Catalogue: context 163840, reasoning, tools
· First seen 2025-09-22

These adoption and deployment observations remain context only. They do not change this model's ModelCap Index score or rank.

BES evidence selected · measured input only

Arena Overall70.1
Arena Coding · Thinking72.0

The bounded 0–100 capability score blends each source’s competitive placement with its published achievement, then combines capability families. Evidence breadth adds a modest uncertainty adjustment; price and popularity are not part of benchmark-led rank.

Deployment readiness

Metadata index

88.5

strong · 98% metadata coverage

Not capability

Artifact reproducibility: 100.0
Access & legal clarity: 100.0
Deployability: 97.5
Evaluation provenance: 46.0

Missing: evaluationProvenance.evaluation-dataset-revisions.

A metadata completeness and deployability index—not a safety certification, quality grade, or production-readiness claim.

Market activity

Venue coverage

Market Gravity: 33.6
OpenRouter weekly popularity: #81 of 298

Market Gravity uses OpenRouter’s full-catalogue popularity order. Sparse Vercel top-ten observations appear only when published and do not affect the score.

Arena preference

Status: Ranked
Source: Arena
Official rank: #119 of 385
Rating: 1415.0
95% confidence interval: 1405–1425
Votes: 4K
Source snapshot: 6 Aug 2026, 13:00 UTC
Category: overall
Identity match confidence: 100%

Market Gravity breakdown

Usage55%: 22.9
Liquidity25%: 70.5
Open reach15%: 12.5
Freshness5%: 29.6

Specification

Context window: 164K tokens
Max output: 33K tokens
Inputs: Text
Outputs: Text
Tokenizer: DeepSeek
Cached input tokens: $0.135 / 1M
First seen on OpenRouter: 22 Sept 2025
Knowledge cutoff: 2025-03-31

Capabilities

Supported: Reasoning
Supported: Tool use
Supported: Structured output
Supported: Response format
Not supported: Moderated

Weights & access

Open weights. Downloadable weights are published under an OSI or free-culture license in the current repository metadata. This describes weight availability, not whether the full training stack qualifies as Open Source AI.

Access: Open weights
Licence: mit
Pinned revision: 19510d6dc61f
Parameters: 684.5B
Architecture: DeepseekV3ForCausalLM
Model type: deepseek_v3
Hugging Face downloads (30d): 17K
Downloads (all time): 222K
Likes: 366
Repository: deepseek-ai/DeepSeek-V3.1-Terminus
Repository updated: 29 Sept 2025, 01:52 UTC
Base lineage: deepseek-ai/DeepSeek-V3.1-Base