← All models

DeepSeek V4.1 Flash (Max)

open weights★ featured
DeepSeek · released 2026-09-10 · 40 offers

Context 1M tokens

Top 5 cheapest providers (Adjusted $/task)

The same list price can give a different adjusted $/task (caching, token efficiency) — click a price for its inputs.

Within the active global provider, residency and confidentiality filters.

#ProviderAdjusted $/task
1Relacevia OpenRouter — Price via OpenRouter: OpenRouter's price for Relace's endpoint (https://openrouter.ai/deepseek/deepseek-v4.1-flash/providers), collected 2026-11-08.
2Morphvia OpenRouter — Price via OpenRouter: OpenRouter's price for Morph's endpoint (https://openrouter.ai/deepseek/deepseek-v4.1-flash/providers), collected 2026-11-08.
3OpenInferencevia OpenRouter — Price via OpenRouter: OpenRouter's price for OpenInference's endpoint (https://openrouter.ai/deepseek/deepseek-v4.1-flash/providers), collected 2026-11-08.
4Wafervia OpenRouter — Price via OpenRouter: OpenRouter's price for Wafer's endpoint (https://openrouter.ai/deepseek/deepseek-v4.1-flash/providers), collected 2026-11-08.
5InferenceNetvia OpenRouter — Price via OpenRouter: OpenRouter's price for InferenceNet's endpoint (https://openrouter.ai/deepseek/deepseek-v4.1-flash/providers), collected 2026-11-08.

Composite

3 of 7 inputs · 2 from the model family6 radar axes: DesignArena's two boards share one

50.7

AA Coding v4.xCoding Agent v1.4AA Intelligence (est.)AA AgenticEpoch ECISoftware ECIDesignArenaAA Intelligence (est.): percentile 89Epoch ECI: percentile 58Software ECI: percentile 39

AA Coding v4.x —Coding Agent v1.4 —AA Intelligence (est.) 29.7AA Agentic —Epoch ECI 156.5Software ECI 152.3DesignArena —/—

Radar: percentile among all models measured on each input; a gap means not measured.

Benchmark sheet

44 of 319 registered benchmark versions · bars show the percentile among all models measured on each benchmark.

19 of 44 values are DeepSeek's own claims (†), not independent measurements; a matching independent result replaces a claim as soon as one exists.

Composite attachments (used in the score, not counted as exact benchmarks):
  • Epoch ECI · attached
  • Software ECI · attached
Compare this model →

Agentic

  • AA-Briefcase821,424 Elo

    AA-Briefcase published 2026-09-10 → · Published board — Tests multi-week professional knowledge-work projects with linked tasks and large source collections.

  • GDPval-AA v2951,632 Elo

    GDPval-AA v2 v2 → · Published board — Tests professional knowledge-work deliverables across occupations using AA's Stirrup harness.

  • Terminal-Bench v4.0 (AA)8626.8%

    Terminal-Bench v4.0 (AA) v4.0 → · Published board — Tests terminal-based work on the 66-task release using mini-SWE-agent v2.4.6.

  • Agent's Last Examdeveloper's claim31.8%† self-reported by the developer

    Agent's Last Exam published 2026-09-10 → · Official scaffold · Vendor-reported by DeepSeek — Agent's Last Exam result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • AutomationBenchdeveloper's claim54.8%† self-reported by the developer

    AutomationBench published 2026-09-10 → · Official scaffold · Vendor-reported by DeepSeek — AutomationBench result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • HLE w/ toolsdeveloper's claim63.9%† self-reported by the developer

    HLE w/ tools published 2026-09-10 → · Vendor-reported by DeepSeek — HLE w/ tools result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • Terminal-Bench 2.1developer's claim90.6%† self-reported by the developer

    Terminal-Bench 2.1 v2.1 → · DeepSeek Harness (Minimal) · Vendor-reported by DeepSeek — Terminal-Bench 2.1 result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • Terminal-Bench 3.0developer's claim30.0%† self-reported by the developer

    Terminal-Bench 3.0 v3.0 → · DeepSeek Harness (Minimal) · Vendor-reported by DeepSeek — Terminal-Bench 3.0 result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • Terminal-Bench 4.0developer's claim31.2%† self-reported by the developer

    Terminal-Bench 4.0 v4.0 → · DeepSeek Harness (Minimal) · Vendor-reported by DeepSeek — Terminal-Bench 4.0 result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • τ²-Bench Airline (OpenRouter run)8776.7%

    τ²-Bench Airline (OpenRouter run) published 2026-09-15 → · Published board — Multi-turn service agents making tool calls under strict policy constraints. (OpenRouter's own reproducible run, as published on openrouter.ai/benchmarks.)

  • LMArena Agent Arena (overall net improvement)424.1%

    LMArena Agent Arena (overall net improvement) published 2026-09-26 → · Published board — Ranking of models as the orchestrator in LMArena’s Agent Mode, scored by the causal net improvement over the average model across behavioural signals from real sessions.

Coding

  • Artificial Analysis Coding Agent Index v2.05647.3%

    Artificial Analysis Coding Agent Index v2.0 v2.0 → · Codex — Measures coding-agent systems on FrontierCode v1.1 Main, Terminal-Bench 4.0 and SWE-Atlas-QnA-AA.

  • SciCode (AA subproblems) v1.0.16551.9%

    SciCode (AA subproblems) v1.0.1 → · Published board — Tests scientific Python programming with scientist-annotated background information.

  • Codeforces (Rating)developer's claim3,471 Elo† self-reported by the developer

    Codeforces (Rating) published 2026-09-10 → · Vendor-reported by DeepSeek — Codeforces (Rating) result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • DeepSWE v1.1developer's claim74.2%† self-reported by the developer

    DeepSWE v1.1 v1.1 → · mini-SWE · Vendor-reported by DeepSeek — DeepSWE v1.1 result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • NL2Repo-Benchdeveloper's claim64.0† self-reported by the developer

    NL2Repo-Bench published 2026-09-10 → · DeepSeek Harness (Minimal) · Vendor-reported by DeepSeek — NL2Repo-Bench result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • ProgramBench (Almost@1)developer's claim20.3%† self-reported by the developer

    ProgramBench (Almost@1) published 2026-09-10 → · DeepSeek Harness (Minimal) · Vendor-reported by DeepSeek — ProgramBench (Almost@1) result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • LMArena WebDev (Code Arena, overall)361,621 Elo

    LMArena WebDev (Code Arena, overall) published 2026-09-26 → · Published board — Community pairwise-vote ranking of models on front-end web development tasks, including agentic multi-step coding, published by LMArena as a Bradley-Terry rating.

Efficiency

Knowledge

Long-context

  • GDP.pdf (AA)5712.8%

    GDP.pdf (AA) published 2026-09-10 → · Published board — Tests professional reasoning over long PDFs with AA document preparation and grading.

  • AA-LCR v1.19884.0%

    AA-LCR v1.1 v1.1 → · Published board — Tests reasoning across multiple long documents with corrected answer keys and grading.

Math

  • MathArena Apexdeveloper's claim65.6%† self-reported by the developer

    MathArena Apex published 2026-09-10 → · Vendor-reported by DeepSeek — MathArena Apex result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • ArXivMath 06/2026 (MathArena)1957.6%

    ArXivMath 06/2026 (MathArena) 2026-06 → · Published board — Research-level math problems with a checkable final answer, drawn from arXiv papers submitted in June 2026; each edition is restricted to papers published within the last month to reduce training-data contamination.

  • ArXivMath 08/2026 (MathArena)043.9%

    ArXivMath 08/2026 (MathArena) 2026-08 → · Published board — Research-level math problems with a checkable final answer, drawn from arXiv papers submitted in August 2026; each edition is restricted to papers published within the last month to reduce training-data contamination.

  • BrokenArXiv 06/2026 (MathArena)2039.4%

    BrokenArXiv 06/2026 (MathArena) 2026-06 → · Published board — Plausible but false proof statements taken from June 2026 arXiv papers; a model scores by refusing to prove them and saying the statement is false as written.

  • BrokenArXiv 08/2026 (MathArena)1354.8%

    BrokenArXiv 08/2026 (MathArena) 2026-08 → · Published board — Plausible but false statements taken from August 2026 arXiv papers; a model scores by refusing to prove them and saying the statement is false as written.

Reasoning

  • GPQA Diamonddeveloper's claim90.9%† self-reported by the developer

    GPQA Diamond published 2026-09-10 → · Vendor-reported by DeepSeek — GPQA Diamond result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • HLEdeveloper's claim36.8%† self-reported by the developer

    HLE published 2026-09-10 → · Vendor-reported by DeepSeek — HLE result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • LiveBench 2026-06-258381.1

    LiveBench 2026-06-25 → · Published board — A periodically refreshed suite of objectively graded tasks with separate category scores and an overall category mean.

Safety/Alignment

  • BullshitBench V1 (clear pushback)7241.8%

    BullshitBench V1 (clear pushback) published 2026-09-10 → · Published board — Whether a model rejects the broken premise of 55 deliberately nonsensical prompts (V1 question set) instead of answering them confidently.

  • BullshitBench V2 (clear pushback)7757.0%

    BullshitBench V2 (clear pushback) published 2026-09-10 → · Published board — Whether a model rejects the broken premise of 100 deliberately nonsensical prompts (V2 question set) instead of answering them confidently.

  • CyberGymdeveloper's claim88.1%† self-reported by the developer

    CyberGym published 2026-09-10 → · Vendor-reported by DeepSeek — CyberGym result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • ExploitGymdeveloper's claim15.3%† self-reported by the developer

    ExploitGym published 2026-09-10 → · Vendor-reported by DeepSeek — ExploitGym result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • SEC-Bench Prodeveloper's claim62.8%† self-reported by the developer

    SEC-Bench Pro published 2026-09-10 → · Claude Code · Vendor-reported by DeepSeek — SEC-Bench Pro result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

Science

  • CritPt (AA)8714.3%

    CritPt (AA) published 2026-09-10 → · Published board — Tests research-level physics reasoning with Python, symbolic and numerical answers.

  • GPQA Diamond (OpenRouter run)8990.2%

    GPQA Diamond (OpenRouter run) published 2026-09-15 → · Published board — Graduate-level science questions that resist retrieval and reward careful reasoning. (OpenRouter's own reproducible run, as published on openrouter.ai/benchmarks.)

Tool-use

  • AutomationBench-AA v1.0.69968.9%

    AutomationBench-AA v1.0.6 → · Published board — Tests multi-app SaaS workflows through REST tools on a held-out AutomationBench split.

Vision

  • MMMU Pro (AA)7577.0%

    MMMU Pro (AA) published 2026-09-10 → · Published board — Tests multimodal understanding using challenging ten-option questions.

  • BabyVision w/ toolsdeveloper's claim89.6%† self-reported by the developer

    BabyVision w/ tools published 2026-09-10 → · Claude Code · Vendor-reported by DeepSeek — BabyVision w/ tools result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • Chartography w/ toolsdeveloper's claim78.9%† self-reported by the developer

    Chartography w/ tools published 2026-09-10 → · Claude Code · Vendor-reported by DeepSeek — Chartography w/ tools result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

  • ZeroBench-main w/ toolsdeveloper's claim49.0%† self-reported by the developer

    ZeroBench-main w/ tools published 2026-09-10 → · Claude Code · Vendor-reported by DeepSeek — ZeroBench-main w/ tools result as reported by DeepSeek in the DeepSeek-V4.1-Flash model card; this registry identity preserves the printed version or the card's publication date when none was stated.

Missing coverage · 275 benchmark versions

No result does not mean a zero, or that the model was never tested. Collection failures and disputed versions retain their distinct status.

Variants / reasoning settings

Artificial Analysis snapshot 2026-11-04 · Intelligence Index v5.0 (est. = estimated from v4, not measured) · Coding Index v4.x frozen 2026-10-27 · Data: Artificial Analysis · Data: Epoch AI (CC BY)

VariantAA CodingAA Intelligence
DeepSeek V4.1 Flash (Max)—29.7est.
DeepSeek V4.1 Flash (Non-reasoning)—12.4est.
Token offers by platform · 34 offers (Adjusted $/task)

Click any underlined price to see how it is estimated and where each input comes from. How we calculate adjusted cost.

34 offers within the active global filters, 6 more outside them; “price n/a” means the provider serves the model but publishes no direct token price we could collect — no other channel's price is shown in its place.

OpenRouter (31)

Prices via OpenRouter — what OpenRouter charges for each provider's endpoint, not the provider's own direct price. Source: openrouter.ai/deepseek/deepseek-v4.1-flash/providers, collected 2026-11-08.

Relace
Morph
OpenInference
Wafer
InferenceNet
InferenceNet
DeepInfra
CoreWeave
Reka
Wafer
Novita
Mancer 2
PrimeIntellect
Fireworks
Phala
DigitalOcean
Makora
Relace
Crusoe
Together
SiliconFlow
Parasail
Fireworks
BaseTen
BaseTen
Venice
Nebius
Google
Modal
Parasail
Fireworks

TensorX (1)

Direct prices from TensorX's own price list. Source: tensorx.ai/pricing/, collected 2026-11-08.

TensorX

Nebius (1)

Direct prices from Nebius's own price list. Source: tokenfactory.nebius.com/models, collected 2026-11-08.

Nebius

T-Systems LLM Hub (1)

Direct prices from T-Systems LLM Hub's own price list. Source: docs.llmhub.t-systems.net/models/llms/, collected 2026-11-05.

T-Systems LLM Hubprice n/a
DeepSeek V4.1 Flash (Max) — benchmarks & cost