RankingEffort benchmarks

DeepSeek V4.1 Flash 1017

Compare models

41 measured results · 5 effort settings · 20 benchmark/harness combinations

Change model

Compare with

Choose a second model and its effort to see both profiles and benchmark differences below.

Compare with

Capability profile

One configuration, seven capabilities

Comparison score · 0–100
DeepSeek V4.1 Flash 1017 · Max capability radarComparison scores from 0 to 100. Unknown capabilities have no point; lines stop at gaps. Hollow points are preliminary. Exact values and support labels follow the chart.AgenticHard reasoningCodingHuman prefKnowledgeMultimodalLong context50100DeepSeek V4.1 Flash 1017 · Max · Agentic: 66.6 · SupportedDeepSeek V4.1 Flash 1017 · Max · Hard reasoning: 53.4 · SupportedDeepSeek V4.1 Flash 1017 · Max · Coding: 75.8 · PreliminaryDeepSeek V4.1 Flash 1017 · Max · Knowledge: 37.3 · SupportedDeepSeek V4.1 Flash 1017 · Max · Multimodal: 47.4 · PreliminaryDeepSeek V4.1 Flash 1017 · Max · Long context: 73.8 · Preliminary

Gaps are unknown, not zero. Hollow points are preliminary.

  • Agentic

    A66.6Supported
  • Hard reasoning

    A53.4Supported
  • Coding

    A75.8Preliminary
  • Human pref

    Unknown
  • Knowledge

    A37.3Supported
  • Multimodal

    A47.4Preliminary
  • Long context

    A73.8Preliminary

Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.

Capability score evidence

DeepSeek V4.1 Flash 1017 · Max: 56.2 overall comparison score. Meets the overall evidence requirements.

6 areas measured. The available task mix affects this estimate; missing areas remain unknown.

Which sources carry the weight?
  • ARC Prize · 28.8% direct comparison weight · independent of the model provider
  • Artificial Analysis · 25.4% direct comparison weight · independent of the model provider
  • Epoch AI · 23.6% direct comparison weight · independent of the model provider
  • Terminal-Bench · 22.3% direct comparison weight · independent of the model provider
Which benchmark families carry the weight?
  • Terminal-Bench · 23.6% direct comparison weight
  • arc-agi-3 · 14.7% direct comparison weight
  • ARC-AGI · 14.1% direct comparison weight
  • GPQA · 6.6% direct comparison weight
  • Aime · 6.6% direct comparison weight
  • Other families (12) · 34.4% direct comparison weight

These shares describe direct comparison weight, not fractions of the final score or confidence. Source balancing does not give every family equal weight or remove correlations between overlapping sources.

How sensitive is this score?

The audit removes one source at a time and refits against the same overall reference panel. These published results use the matching dataset audited 2026-11-22.

Both ranks use the same surviving configurations for each removal. The cohort starts with 602 configurations supported in the original audit, including separate effort settings. These are diagnostic positions, not the default leaderboard ranks. Eligibility is not rechecked after removal.

What changes when a source is removed?
Removed sourceRank before → afterRank movementCohort remaining / lost supportScore after (change)
Artificial Analysis80 → 82Down 2230 remaining / 372 lost54.8 (-1.4 points)
ARC Prize92 → 85Up 7602 remaining / 0 lost63.8 (+7.6 points)
Datacurve92 → 90Up 2602 remaining / 0 lost51.4 (-4.8 points)
LiveCodeBench92 → 92No movement602 remaining / 0 lost56.2 (-0.0 points)
Scale92 → 93Down 1602 remaining / 0 lost56.8 (+0.7 points)
Terminal-Bench92 → 104Down 12602 remaining / 0 lost51.1 (-5.1 points)
Cognition92 → 85Up 7602 remaining / 0 lost54.6 (-1.5 points)
LiveBench92 → 91Up 1601 remaining / 1 lost57.7 (+1.5 points)
SWE-rebench92 → 91Up 1602 remaining / 0 lost55.8 (-0.4 points)
Epoch AI92 → 82Up 10588 remaining / 14 lost60.2 (+4.0 points)
Vals AI92 → 92No movement601 remaining / 1 lost55.6 (-0.5 points)
Provider reports · Anthropic92 → 92No movement602 remaining / 0 lost56.1 (-0.1 points)
Provider reports · Moonshot AI92 → 92No movement602 remaining / 0 lost55.4 (-0.7 points)
Provider reports · Meta92 → 92No movement602 remaining / 0 lost56.1 (-0.0 points)
Provider reports · Mistral92 → 92No movement602 remaining / 0 lost56.2 (+0.0 points)
Provider reports · OpenAI92 → 92No movement602 remaining / 0 lost56.2 (+0.0 points)
All provider reports92 → 93Down 1602 remaining / 0 lost55.4 (-0.8 points)

Lost support means a configuration no longer has a comparison path to the complete reference panel. Its rank is unavailable. A surviving path alone does not establish leaderboard eligibility.

Overlapping sources may remain correlated even with source balancing. Leave-one-source-out movement is a sensitivity diagnostic, not a spread of source-specific ranks or a confidence measure. Scoring methodology · Dated configuration audit

Every benchmark family
Removed evidenceRecomputed scoreChange
terminal-bench51.2-4.9
arc-agi-361.1+4.9
deepswe51.4-4.7
simpleqa60.1+3.9
aime54.5-1.7
frontiercode54.6-1.5
mmmu-pro57.4+1.2
gpqa55.3-0.9
epoch-game-puzzles55.4-0.7
frontiermath55.5-0.7
gdp-pdf56.8+0.6
hle56.8+0.6
briefcase55.6-0.6
automationbench55.6-0.5
enigma-eval56.7+0.5
mmlu-pro56.7+0.5
gdpval55.7-0.4
livebench-instructions56.6+0.4
omniscience56.6+0.4
swe-rebench55.8-0.4
multi-swe-bench55.8-0.4
swe-atlas-qna55.8-0.3
livebench-coding55.9-0.3
livebench-language56.4+0.3
ifbench55.9-0.3
critpt56.4+0.3
vibe-code55.9-0.2
swe-atlas-test-writing55.9-0.2
vals-legal-research56.0-0.2
tau-bench56.0-0.2
livebench-data56.4+0.2
arc-agi56.3+0.2
livebench-reasoning56.0-0.1
vals-tax-agent56.1-0.1
swe-bench-pro56.1-0.1
apex-agents56.3+0.1
vals-code-migration56.1-0.1
harvey56.2+0.1
livebench-math56.1-0.1
aa-lcr56.1-0.1
vals-excel-modeling56.1-0.1
terminal-bench-science56.1-0.1
vals-finance-agent56.1-0.1
enterprise-ops56.2+0.0
livecodebench56.1-0.0
swe-atlas-refactoring56.2+0.0
itbench56.2+0.0
math50056.2-0.0
gmmlu56.2-0.0
analyst-agent56.2+0.0
What evidence is missing from the fit?

Across all collected settings for this model: 41 observations, 41 contributing, 0 matched without graph weight and 0 excluded. Counts do not establish rank eligibility.

    Inspect every observation and exclusion →

    Benchmark results

    These are published benchmark scores. Supported effort estimates also appear on the Capability leaderboard. Results from different harnesses occupy different rows. Missing cells stay unknown; no scores are copied between effort levels.

    Reviewed retained observations are in use. October 2026 Real-SWE board has no 95% confidence whisker, so these scores stay at the 2026-10-18 retrieval. Original source retrieval 2026-10-18; source updated 2026-10-18.

    Benchmark / sourceLowHighMaxNoneNon-reasoning
    AA-Briefcase EloArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    AA-Briefcase v1.1 is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better · Evaluation results measured independently by Artificial Analysis

    Unit: elo. Source collected 2026-11-22.

    Not reported
    1,415.98 Elo
    Reported settings

    DeepSeek V4.1 Flash 1017 (High); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-high

    1,570.42 Elo
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    1,425.93 Elo
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    AA-LCR v1.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    82.67%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    57.33%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    AutomationBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Share of task objectives completed with no guardrail violations · Higher is better · Benchmark developed by Zapier · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    Not reported
    68.49%
    Reported settings

    DeepSeek V4.1 Flash 1017 (High); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-high

    66.64%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    48.78%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    CritPtArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Benchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    24%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    0.29%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    GDP.pdf: All-pass RateArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Share of attempts where every atomic criterion passed · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    Not reported
    14.2%
    Reported settings

    DeepSeek V4.1 Flash 1017 (High); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-high

    17.2%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    4.4%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    GDPval-AA v2.1 LeaderboardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Elo rating for performance on real-world work tasks · Anchored to DeepSeek V4.1 Flash (max) at 1600 · Higher is better · Evaluation results measured independently by Artificial Analysis

    Unit: elo. Source collected 2026-11-22.

    Not reportedNot reported
    1,745.77 Elo
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    1,317.14 Elo
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    Humanity's Last ExamArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    43.33%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    11.54%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    MLCR-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Share of responses judged complete with no claim contradicting the record · Expert + compound (hard) sets · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    26.67%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reportedNot reported
    MMMU-ProArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings
    Benchmark details

    Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    Not reported
    79.71%
    Reported settings

    DeepSeek V4.1 Flash 1017 (High); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-high

    80.64%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    69.19%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    AA-Omniscience AccuracyArtificial Analysis · 3000 questions; 1 attempt; correct-answer share; GPT-5.6 Luna medium grader
    Benchmark details

    AA-Omniscience v1.1 Accuracy (higher is better) measures the proportion of correctly answered questions out of all questions, regardless of whether the model chooses to answer · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    Not reported
    38.93%
    Reported settings

    DeepSeek V4.1 Flash 1017 (High); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-high

    39.6%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    21.87%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    Terminal-Bench-Science 0.1Artificial Analysis · mini-swe-agent; Terminal-Bench-Science 0.1.0; 70 tasks; pass@1 averaged over 3 repeats
    Benchmark details

    Share of tasks passed (pass@1), averaged over three repeats · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    Not reported
    10.48%
    Reported settings

    DeepSeek V4.1 Flash 1017 (High); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-high

    12.86%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    2.38%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    Terminal-Bench v4.0Artificial Analysis · mini-SWE-agent v2.4.6; 66 tasks; pass@1 averaged over 3 repeats; 500 steps; task timeout up to 8 hours
    Benchmark details

    Benchmark developed by the Laude Institute, Stanford, and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis

    Unit: percent. Source collected 2026-11-22.

    Not reported
    28.79%
    Reported settings

    DeepSeek V4.1 Flash 1017 (High); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-high

    33.84%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017

    Not reported
    9.09%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017-non-reasoning

    ARC-AGI-2ARC Prize · v2 Semi-Private
    Benchmark details

    Original public board. Configuration and task coverage may differ; inspect source.

    Unit: percent. Source collected 2026-11-22.

    73.61%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Low); Source model ID: deepseek-v4-1-flash-1017-low; 120 tasks

    68.61%
    Reported settings

    DeepSeek V4.1 Flash 1017 (High); Source model ID: deepseek-v4-1-flash-1017-high; 120 tasks

    78.06%
    Reported settings

    DeepSeek V4.1 Flash 1017 (Max); Source model ID: deepseek-v4-1-flash-1017-max; 120 tasks

    2.08%
    Reported settings

    DeepSeek V4.1 Flash 1017 (None); Source model ID: deepseek-v4-1-flash-1017-none; 120 tasks

    Not reported
    ARC-AGI-3 · provider-adapterARC Prize · Provider Adapter
    Benchmark details

    ARC-AGI-3; Provider Adapter. Open each result for its exact effort, task subset, metric and protocol details.

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    5.81%
    Reported settings

    ARC-AGI-3 Semi-Private verified summary; Provider Adapter harness; source group deepseek-v4-1-flash-1017; variant provider-adapter-max; public-demo task scores are excluded

    Not reportedNot reported
    ARC-AGI-3 · standardARC Prize · Standard
    Benchmark details

    ARC-AGI-3; Standard. Open each result for its exact effort, task subset, metric and protocol details.

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    0.22%
    Reported settings

    ARC-AGI-3 Semi-Private verified summary; Standard harness; source group deepseek-v4-1-flash-1017; variant max; public-demo task scores are excluded

    Not reportedNot reported
    Terminal-Bench 4.0 · claudecodeTerminal-Bench · Claude Code
    Benchmark details

    Terminal-Bench 4.0; Claude Code. Open each result for its exact effort, task subset, metric and protocol details.

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    40%
    Reported settings

    Terminal-Bench 4.0.0; 66 tasks; Claude Code; effort max; reported margin ±3.7 percentage points, confidence level unspecified; release date 2026-10-17; board checked 2026-11-18

    Not reportedNot reported
    Epoch · FrontierMath Tiers 1–3 · v2.0.0Epoch AI · Epoch AI Inspect; task 2.0.0; scorer verification_code
    Benchmark details

    Epoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY. FrontierMath was developed with OpenAI funding and unequal access to part of the question set; this result is from Epoch's evaluation.

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    71.23%
    Reported settings

    Epoch AI Inspect; task 2.0.0; scorer verification_code; model=deepseek-v4.1-flash-1017_max; run=UnaMiD4Yy8W4WzyeSRwybr; mean accuracy; https://epoch.ai/data/benchmarks.csv

    Not reportedNot reported
    Epoch · GPQA diamond · v1.0.11Epoch AI · Epoch AI Inspect; task 1.0.11; scorer choice
    Benchmark details

    Epoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY.

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    93.18%
    Reported settings

    Epoch AI Inspect; task 1.0.11; scorer choice; model=deepseek-v4.1-flash-1017_max; run=ZK85PJtsCSeQJgV8dA4R4G; mean accuracy; https://epoch.ai/data/benchmarks.csv

    Not reportedNot reported
    Epoch · OTIS Mock AIME 2024-2025 · v1.1.6Epoch AI · Epoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match
    Benchmark details

    Epoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY.

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    99.44%
    Reported settings

    Epoch AI Inspect; task 1.1.6; scorer model_extracted_exact_match; model=deepseek-v4.1-flash-1017_max; run=chayEvHvpt2Wcif7pg64hF; mean accuracy; https://epoch.ai/data/benchmarks.csv

    Not reportedNot reported
    Epoch · SimpleQA Verified · v1.2.0Epoch AI · Epoch AI Inspect; task 1.2.0; scorer simpleqa_scorer
    Benchmark details

    Epoch AI internal evaluation. Mean accuracy under the published setting; exact task version and scorer stay separate. Related tasks share a family budget. Data by Epoch AI, CC BY.

    Unit: percent. Source collected 2026-11-22.

    Not reportedNot reported
    42.5%
    Reported settings

    Epoch AI Inspect; task 1.2.0; scorer simpleqa_scorer; model=deepseek-v4.1-flash-1017_max; run=Ei4XWaCH5drYUnDRo2MYCW; mean accuracy; https://epoch.ai/data/benchmarks.csv

    Not reportedNot reported
    What these comparisons mean

    Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.

    These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.

    Download all effort benchmark results · 10,184 results across 381 models · Collected 2026-11-22