Language modelActive

DeepSeek-V4-Pro

DeepSeek

Released: April 24, 2026
Data date: September 8, 2026

Benchmark-Meta-Score: Unavailable
Minimum coverage not met
Speed: 70.8 tokens/s
Output throughput
Price: $0.66 · $1.98
Input · Output / 1M tokens (Off-Peak)
Input: Text
Output: Text

New in DeepSeek-V4-Pro is the combination of an open V4-Pro checkpoint and a 1,000,000-token context window. DeepSeek describes 1.6 trillion total parameters and 49 billion active parameters.

Since September 14, 2026, DeepSeek has temporarily routed the deepseek-v4-pro alias to DeepSeek-V4.1-Flash until V4.1-Pro launches. The released weights remain separate.

Specifications and access

Specification	Value and source
Weights	Open weightsSource
Access	API, LocalSource
Parameters	1.6TSource
Active parameters	49BSource
Context window	1,000,000 tokensSource
Notes	1.6T total, 49B active, 1M context, CSA+HCA attention, MIT-licensedSource
License	MITSource
Parameters	1,600 BSource
Active parameters	49 BSource
Architecture	MoE with CSA/HCA and mHCSource
Training tokens	>32TSource
Input	TextSource
Output	TextSource
Reasoning mode	non-think, think-high, think-maxSource
Tool calling	Yes (Native tool use is documented for V4 modes in official release documentation)Source
API model ID	deepseek-v4-pro (Pricing page updated alias: V4-Pro-0813)Source

Benchmark meta-analysis

Overall and area scores

These Gradually indexes place selected published measurements on a 0 to 100 scale. They are not accuracy percentages and do not guarantee results in your application.

Retrieval dates of the sources used 2026-10-15 - 2026-10-19

Overall index

Not enough coverage for a score yet

provisional

Published configurations differ within this comparison cohort. Treat the index as provisional.

Panel-weight coverage: 40 %
Test families: 33 / 71
Sources: 3
Mathematical bounds with missing values: 16.9-76.8 / 100

Coding

Not enough coverage for a score yet

provisional

Published configurations differ within this comparison cohort. Treat the index as provisional.

Panel-weight coverage: 40 %
Test families: 7 / 15
Sources: 2
Mathematical bounds with missing values: 20.4-80.4 / 100

Reasoning

Index 45.2 / 100

provisional

Published configurations differ within this comparison cohort. Treat the index as provisional.

Panel-weight coverage: 64 %
Test families: 5 / 7
Sources: 3
Mathematical bounds with missing values: 29.1-64.8 / 100

Mathematics

Index 16.6 / 100

provisional

Published configurations differ within this comparison cohort. Treat the index as provisional.

Panel-weight coverage: 60 %
Test families: 3 / 5
Sources: 2
Mathematical bounds with missing values: 10.0-50.0 / 100

Knowledge

Index 51.4 / 100

provisional

Published configurations differ within this comparison cohort. Treat the index as provisional.

Panel-weight coverage: 100 %
Test families: 3 / 3
Sources: 3
Mathematical bounds with missing values: 51.4-51.4 / 100

Agents

Not enough coverage for a score yet

provisional

Published configurations differ within this comparison cohort. Treat the index as provisional.

Panel-weight coverage: 33 %
Test families: 4 / 12
Sources: 2
Mathematical bounds with missing values: 21.8-88.5 / 100

Finance

Index 42.3 / 100

provisional

Published configurations differ within this comparison cohort. Treat the index as provisional.

Panel-weight coverage: 71 %
Test families: 5 / 7
Sources: 1
Mathematical bounds with missing values: 30.2-58.8 / 100

Legal

Index 40.7 / 100

provisional

Published configurations differ within this comparison cohort. Treat the index as provisional.

Panel-weight coverage: 75 %
Test families: 3 / 4
Sources: 1
Mathematical bounds with missing values: 30.5-55.5 / 100

Health

Index 26.7 / 100

provisional

Published configurations differ within this comparison cohort. Treat the index as provisional.

Panel-weight coverage: 67 %
Test families: 2 / 3
Sources: 1
Mathematical bounds with missing values: 17.8-51.1 / 100

Productivity

Not enough coverage for a score yet

provisional

Published configurations differ within this comparison cohort. Treat the index as provisional.

Panel-weight coverage: 50 %
Test families: 1 / 2
Sources: 1
Mathematical bounds with missing values: 25.0-75.0 / 100

Documented method · Included test families and measurements

Method version: benchmark-meta-score-v1

The index includes active measurements from reviewed primary sources when at least five models are compared in the same documented comparison slice across dataset, harness, agent, tools, and trial count. Each benchmark series and evaluator contributes one shared setup slice from the newest capture; when dates match, the wider slice is used. When one source and series include several explicitly documented reasoning tiers, the highest documented tier (none through max) is selected, never the best result value. The explicit source label “Non-reasoning” is treated as none. When selecting between multiple results, unknown or ambiguous tiers exclude the model from that series. Unambiguous single configurations do not require a tier ordering. The best individual result for each model is not selected. Equal raw values share the same rank percentile. Multiple series from the same evaluator within a family receive equal weight; evaluators within a family, families within an area, and areas in the overall index then receive equal weight. Unmeasured series remain visible in coverage and bounds. An overall score requires at least four areas, six test families, and two sources. An area score requires at least two test families. Every overall and area score also requires at least 50% of its panel weight.

Published model-specific parameters can differ. This is not a controlled experiment; comparisons with such differences are marked provisional.

Unmeasured portions are not entered as zero. They reduce coverage and contribute only to the displayed bounds. Scores with different coverage are therefore not equally complete.

The bounds hold observed percentiles fixed and allow missing contributions anywhere from 0 to 100. They are not confidence intervals or predictions of future recalculations.

The source count is not a count of independent experiments.

OECD handbook on composite indicators

All 33 test families and 12 of 38 raw measurements are listed here. The complete calculation for every model is available in the JSON export.

Test family: AIME
Mathematics · Family index: 31.8 / 100
Nominal family weight in the overall panel: 1.43 % · Coverage of this test family: 100 %
- Epoch AI benchmark run dataset snapshot
  Published configurations differ within this cohort.
  Measurement:
  epoch-otis-mock-aime-2024-2025
  Series weight within the test family:
  100 %
  Nominal contribution to the overall panel:
  1.43 %
  Raw value:
  0.9666666666666667 ratio
  Metric:
  model_extracted_exact_match
  Version:
  1.1.6
  Participating models:
  23
  Rank percentile within the cohort:
  31.8 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  epoch-score-design-v1:["OTIS Mock AIME 2024-2025","OTIS Mock AIME 2024-2025","1.1.6","model_extracted_exact_match"]
  Reasoning tier:
  max
  Documented comparison setup
  Publication cohort:
  epoch-score-design-v1:["OTIS Mock AIME 2024-2025","OTIS Mock AIME 2024-2025","1.1.6","model_extracted_exact_match"]
  Dataset type:
  not-reported-in-captured-run-csv
  Setup and model parameters
  Published reasoning tier: max. The capture does not report other inference parameters or a compute budget.
  Independent evaluation. Epoch run alias: deepseek-v4-pro_max; Epoch model label: DeepSeek-V4-Pro; Display label field: DeepSeek v4 Pro (max); Reasoning setting encoded in the alias: max. Task ID: OTIS Mock AIME 2024-2025. Task version: 1.1.6. Run ID: HjTDENSpy9sancDzr7pccY. Started at 2026-06-17T00:34:52.000Z. Original task label: OTIS Mock AIME 2024-2025. The published row reports mean_score=0.9666666666666667 and stderr=0.02002208376727653. The CSV has no sample-count column, so none is inferred. Standard error is not presented as a confidence interval. Meta-score selection: Per model, the highest documented reasoning tier is selected within the same Epoch source test design, original task, version, and scorer, independently of the measured value. This derived score copy does not claim matched compute budgets; the capture publishes neither a common budget nor all inference parameters. The raw model-profile observation remains unchanged. Selected reasoning tier: max.
Test family: AA-Briefcase
Agents · Family index: 67.1 / 100
Nominal family weight in the overall panel: 0.6 % · Coverage of this test family: 100 %
- Artificial Analysis
  Published configurations differ within this cohort.
  Measurement:
  artificial-analysis-aa-briefcase
  Series weight within the test family:
  100 %
  Nominal contribution to the overall panel:
  0.6 %
  Raw value:
  1,256.94 Elo
  Metric:
  AA-Briefcase Elo
  Version:
  v1.1
  Participating models:
  83
  Rank percentile within the cohort:
  67.1 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  artificial-analysis-aa-briefcase:v1.1:v4.3.2:reasoning:tokens=2000:suffix=Max
  Reasoning tier:
  Max
  Documented comparison setup
  Publication cohort:
  ["artificial-analysis-aa-briefcase","v1.1","aa-model-page-2026-10-19-8a022369738c"]
  Dataset type:
  artificial-analysis-evaluation
  Setup and model parameters
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Reasoning: reasoning enabled. Reasoning effort: Max
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Source model ID: artificial-analysis/83173329-c09a-41f1-a028-a282a5f908d7/deepseek-v4-pro/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Model snapshot: DeepSeek V4 Pro 0813 (Max). Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Reasoning: reasoning enabled. Reasoning effort: Max. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-10-19-8a022369738c. Source payload hash: 8a022369738cbe2a0e2f9bf486a2882865dbb0c6f1396dfec84059a802560e3c. Parser version: artificial-analysis-rsc-models-v2. Imported at: 2026-10-19
Test family: AA-LCR
Reasoning · Family index: 69.5 / 100
Nominal family weight in the overall panel: 1.02 % · Coverage of this test family: 100 %
- Artificial Analysis
  Published configurations differ within this cohort.
  Measurement:
  artificial-analysis-aa-lcr
  Series weight within the test family:
  100 %
  Nominal contribution to the overall panel:
  1.02 %
  Raw value:
  80.3333333333333 %
  Metric:
  Pass rate
  Version:
  v1.1
  Participating models:
  83
  Rank percentile within the cohort:
  69.5 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  artificial-analysis-aa-lcr:v1.1:v4.3.2:reasoning:tokens=2000:suffix=Max
  Reasoning tier:
  Max
  Documented comparison setup
  Publication cohort:
  ["artificial-analysis-aa-lcr","v1.1","aa-model-page-2026-10-19-8a022369738c"]
  Dataset type:
  artificial-analysis-evaluation
  Setup and model parameters
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Reasoning: reasoning enabled. Reasoning effort: Max
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Source model ID: artificial-analysis/83173329-c09a-41f1-a028-a282a5f908d7/deepseek-v4-pro/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Model snapshot: DeepSeek V4 Pro 0813 (Max). Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Reasoning: reasoning enabled. Reasoning effort: Max. Score display: 80.3%. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-10-19-8a022369738c. Source payload hash: 8a022369738cbe2a0e2f9bf486a2882865dbb0c6f1396dfec84059a802560e3c. Parser version: artificial-analysis-rsc-models-v2. Imported at: 2026-10-19
Test family: AutomationBench-AA
Agents · Family index: 76.8 / 100
Nominal family weight in the overall panel: 0.6 % · Coverage of this test family: 100 %
- Artificial Analysis
  Published configurations differ within this cohort.
  Measurement:
  artificial-analysis-automationbench-aa
  Series weight within the test family:
  100 %
  Nominal contribution to the overall panel:
  0.6 %
  Raw value:
  56.711655463440025 %
  Metric:
  Objectives completed without guardrail violations
  Version:
  unspecified
  Participating models:
  83
  Rank percentile within the cohort:
  76.8 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  artificial-analysis-automationbench-aa:unspecified:v4.3.2:reasoning:tokens=2000:suffix=Max
  Reasoning tier:
  Max
  Documented comparison setup
  Publication cohort:
  ["artificial-analysis-automationbench-aa","unspecified","aa-model-page-2026-10-19-8a022369738c"]
  Dataset type:
  artificial-analysis-evaluation
  Setup and model parameters
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Reasoning: reasoning enabled. Reasoning effort: Max
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Source model ID: artificial-analysis/83173329-c09a-41f1-a028-a282a5f908d7/deepseek-v4-pro/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Model snapshot: DeepSeek V4 Pro 0813 (Max). Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Reasoning: reasoning enabled. Reasoning effort: Max. Score display: 56.7%. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-10-19-8a022369738c. Source payload hash: 8a022369738cbe2a0e2f9bf486a2882865dbb0c6f1396dfec84059a802560e3c. Parser version: artificial-analysis-rsc-models-v2. Imported at: 2026-10-19
Test family: GDP.pdf
Reasoning · Family index: 53.7 / 100
Nominal family weight in the overall panel: 1.02 % · Coverage of this test family: 100 %
- Artificial Analysis
  Published configurations differ within this cohort.
  Measurement:
  artificial-analysis-gdp-pdf
  Series weight within the test family:
  100 %
  Nominal contribution to the overall panel:
  1.02 %
  Raw value:
  11.4 %
  Metric:
  All-pass rate
  Version:
  unspecified
  Participating models:
  83
  Rank percentile within the cohort:
  53.7 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  artificial-analysis-gdp-pdf:unspecified:v4.3.2:reasoning:tokens=2000:suffix=Max
  Reasoning tier:
  Max
  Documented comparison setup
  Publication cohort:
  ["artificial-analysis-gdp-pdf","unspecified","aa-model-page-2026-10-19-8a022369738c"]
  Dataset type:
  artificial-analysis-evaluation
  Setup and model parameters
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Reasoning: reasoning enabled. Reasoning effort: Max
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Source model ID: artificial-analysis/83173329-c09a-41f1-a028-a282a5f908d7/deepseek-v4-pro/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Model snapshot: DeepSeek V4 Pro 0813 (Max). Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Reasoning: reasoning enabled. Reasoning effort: Max. Score display: 11.4%. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-10-19-8a022369738c. Source payload hash: 8a022369738cbe2a0e2f9bf486a2882865dbb0c6f1396dfec84059a802560e3c. Parser version: artificial-analysis-rsc-models-v2. Imported at: 2026-10-19
Test family: GDPval-AA
Agents · Family index: 73.2 / 100
Nominal family weight in the overall panel: 0.6 % · Coverage of this test family: 100 %
- Artificial Analysis
  Published configurations differ within this cohort.
  Measurement:
  artificial-analysis-gdpval-aa
  Series weight within the test family:
  100 %
  Nominal contribution to the overall panel:
  0.6 %
  Raw value:
  1,450.47 Elo
  Metric:
  GDPval-AA Elo
  Version:
  v2.1
  Participating models:
  83
  Rank percentile within the cohort:
  73.2 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  artificial-analysis-gdpval-aa:v2.1:v4.3.2:reasoning:tokens=2000:suffix=Max
  Reasoning tier:
  Max
  Documented comparison setup
  Publication cohort:
  ["artificial-analysis-gdpval-aa","v2.1","aa-model-page-2026-10-19-8a022369738c"]
  Dataset type:
  artificial-analysis-evaluation
  Setup and model parameters
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Reasoning: reasoning enabled. Reasoning effort: Max
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Source model ID: artificial-analysis/83173329-c09a-41f1-a028-a282a5f908d7/deepseek-v4-pro/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Model snapshot: DeepSeek V4 Pro 0813 (Max). Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Reasoning: reasoning enabled. Reasoning effort: Max. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-10-19-8a022369738c. Source payload hash: 8a022369738cbe2a0e2f9bf486a2882865dbb0c6f1396dfec84059a802560e3c. Parser version: artificial-analysis-rsc-models-v2. Imported at: 2026-10-19
Test family: AA-Omniscience
Knowledge · Family index: 54.9 / 100
Nominal family weight in the overall panel: 2.38 % · Coverage of this test family: 100 %
- Artificial Analysis
  Published configurations differ within this cohort.
  Measurement:
  artificial-analysis-omniscience
  Series weight within the test family:
  100 %
  Nominal contribution to the overall panel:
  2.38 %
  Raw value:
  0.8333333333333334 points
  Metric:
  AA-Omniscience Index
  Version:
  unspecified
  Participating models:
  83
  Rank percentile within the cohort:
  54.9 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  artificial-analysis-omniscience:unspecified:v4.3.2:reasoning:tokens=2000:suffix=Max
  Reasoning tier:
  Max
  Documented comparison setup
  Publication cohort:
  ["artificial-analysis-omniscience","unspecified","aa-model-page-2026-10-19-8a022369738c"]
  Dataset type:
  artificial-analysis-evaluation
  Setup and model parameters
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Reasoning: reasoning enabled. Reasoning effort: Max
  Exact AA configuration: DeepSeek V4 Pro 0813 (Max) (release deepseek-v4-pro, row deepseek-v4-pro, UUID 83173329-c09a-41f1-a028-a282a5f908d7, reasoning=true, source reasoning-token configuration budget=2000, suffix=Max). This configured value is not measured token consumption.. Source model ID: artificial-analysis/83173329-c09a-41f1-a028-a282a5f908d7/deepseek-v4-pro/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Model snapshot: DeepSeek V4 Pro 0813 (Max). Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Reasoning: reasoning enabled. Reasoning effort: Max. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-10-19-8a022369738c. Source payload hash: 8a022369738cbe2a0e2f9bf486a2882865dbb0c6f1396dfec84059a802560e3c. Parser version: artificial-analysis-rsc-models-v2. Imported at: 2026-10-19
Test family: Chess Puzzles
Reasoning · Family index: 0.0 / 100
Nominal family weight in the overall panel: 1.02 % · Coverage of this test family: 100 %
- Epoch AI benchmark run dataset snapshot
  Published configurations differ within this cohort.
  Measurement:
  epoch-chess-puzzles
  Series weight within the test family:
  100 %
  Nominal contribution to the overall panel:
  1.02 %
  Raw value:
  0.19 ratio
  Metric:
  exact
  Version:
  2.0.0
  Participating models:
  22
  Rank percentile within the cohort:
  0.0 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  epoch-score-design-v1:["Chess Puzzles","Chess Puzzles","2.0.0","exact"]
  Reasoning tier:
  max
  Documented comparison setup
  Publication cohort:
  epoch-score-design-v1:["Chess Puzzles","Chess Puzzles","2.0.0","exact"]
  Dataset type:
  not-reported-in-captured-run-csv
  Setup and model parameters
  Published reasoning tier: max. The capture does not report other inference parameters or a compute budget.
  Independent evaluation. Epoch run alias: deepseek-v4-pro_max; Epoch model label: DeepSeek-V4-Pro; Display label field: DeepSeek v4 Pro (max); Reasoning setting encoded in the alias: max. Task ID: Chess Puzzles. Task version: 2.0.0. Run ID: DSyqeLiA9LLZF4iaxsTVHL. Started at 2026-10-13T13:38:19.000Z. Original task label: Chess Puzzles. The published row reports mean_score=0.19 and stderr=0.039427724440366255. The CSV has no sample-count column, so none is inferred. Standard error is not presented as a confidence interval. Meta-score selection: Per model, the highest documented reasoning tier is selected within the same Epoch source test design, original task, version, and scorer, independently of the measured value. This derived score copy does not claim matched compute budgets; the capture publishes neither a common budget nor all inference parameters. The raw model-profile observation remains unchanged. Selected reasoning tier: max.
Test family: SimpleQA Verified
Knowledge · Family index: 27.8 / 100
Nominal family weight in the overall panel: 2.38 % · Coverage of this test family: 100 %
- Epoch AI benchmark run dataset snapshot
  Published configurations differ within this cohort.
  Measurement:
  epoch-simpleqa-verified
  Series weight within the test family:
  100 %
  Nominal contribution to the overall panel:
  2.38 %
  Raw value:
  0.46993987975951906 ratio
  Metric:
  simpleqa_scorer
  Version:
  1.2.0
  Participating models:
  19
  Rank percentile within the cohort:
  27.8 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  epoch-score-design-v1:["SimpleQA Verified","SimpleQA Verified (anti-abstention)","1.2.0","simpleqa_scorer"]
  Reasoning tier:
  max
  Documented comparison setup
  Publication cohort:
  epoch-score-design-v1:["SimpleQA Verified","SimpleQA Verified (anti-abstention)","1.2.0","simpleqa_scorer"]
  Dataset type:
  not-reported-in-captured-run-csv
  Setup and model parameters
  Published reasoning tier: max. The capture does not report other inference parameters or a compute budget.
  Independent evaluation. Epoch run alias: deepseek-v4-pro_max; Epoch model label: DeepSeek-V4-Pro; Display label field: DeepSeek v4 Pro (max); Reasoning setting encoded in the alias: max. Task ID: SimpleQA Verified. Task version: 1.2.0. Run ID: Sq9xqkVTmFvELpQy9KH5TM. Started at 2026-08-27T19:30:29.000Z. Original task label: SimpleQA Verified (anti-abstention). The published row reports mean_score=0.46993987975951906 and stderr=0.015806515332355317. The CSV has no sample-count column, so none is inferred. Standard error is not presented as a confidence interval. Meta-score selection: Per model, the highest documented reasoning tier is selected within the same Epoch source test design, original task, version, and scorer, independently of the measured value. This derived score copy does not claim matched compute budgets; the capture publishes neither a common budget nor all inference parameters. The raw model-profile observation remains unchanged. Selected reasoning tier: max.
Test family: FrontierMath
Mathematics · Family index: 0.0 / 100
Nominal family weight in the overall panel: 1.43 % · Coverage of this test family: 100 %
- Epoch AI benchmark run dataset snapshot
  Published configurations differ within this cohort.
  Measurement:
  epoch-frontiermath-tier-4-v2-private
  Series weight within the test family:
  50 %
  Nominal contribution to the overall panel:
  0.71 %
  Raw value:
  0.024390243902439025 ratio
  Metric:
  verification_code
  Version:
  2.0.0
  Participating models:
  19
  Rank percentile within the cohort:
  0.0 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  epoch-score-design-v1:["FrontierMath-Tier-4-v2-Private","FrontierMath-Tier-4-v2-Private","2.0.0","verification_code"]
  Reasoning tier:
  max
  Documented comparison setup
  Publication cohort:
  epoch-score-design-v1:["FrontierMath-Tier-4-v2-Private","FrontierMath-Tier-4-v2-Private","2.0.0","verification_code"]
  Dataset type:
  not-reported-in-captured-run-csv
  Setup and model parameters
  Published reasoning tier: max. The capture does not report other inference parameters or a compute budget.
  Independent evaluation. Epoch run alias: deepseek-v4-pro_max; Epoch model label: DeepSeek-V4-Pro; Display label field: DeepSeek v4 Pro (max); Reasoning setting encoded in the alias: max. Task ID: FrontierMath-Tier-4-v2-Private. Task version: 2.0.0. Run ID: RAZ6VMfGtQSE8EvUZRRz4Q. Started at 2026-06-17T02:30:14.000Z. Original task label: FrontierMath-Tier-4-v2-Private. The published row reports mean_score=0.024390243902439025 and stderr=0.024390243902439025. The CSV has no sample-count column, so none is inferred. Standard error is not presented as a confidence interval. Meta-score selection: Per model, the highest documented reasoning tier is selected within the same Epoch source test design, original task, version, and scorer, independently of the measured value. This derived score copy does not claim matched compute budgets; the capture publishes neither a common budget nor all inference parameters. The raw model-profile observation remains unchanged. Selected reasoning tier: max.
- Epoch AI benchmark run dataset snapshot
  Published configurations differ within this cohort.
  Measurement:
  epoch-frontiermath-tiers-1-3-v2-private
  Series weight within the test family:
  50 %
  Nominal contribution to the overall panel:
  0.71 %
  Raw value:
  0.45263157894736844 ratio
  Metric:
  verification_code
  Version:
  2.0.0
  Participating models:
  23
  Rank percentile within the cohort:
  0.0 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  epoch-score-design-v1:["FrontierMath-Tiers-1-3-v2-Private","FrontierMath-Tiers-1-3-v2-Private","2.0.0","verification_code"]
  Reasoning tier:
  max
  Documented comparison setup
  Publication cohort:
  epoch-score-design-v1:["FrontierMath-Tiers-1-3-v2-Private","FrontierMath-Tiers-1-3-v2-Private","2.0.0","verification_code"]
  Dataset type:
  not-reported-in-captured-run-csv
  Setup and model parameters
  Published reasoning tier: max. The capture does not report other inference parameters or a compute budget.
  Independent evaluation. Epoch run alias: deepseek-v4-pro_max; Epoch model label: DeepSeek-V4-Pro; Display label field: DeepSeek v4 Pro (max); Reasoning setting encoded in the alias: max. Task ID: FrontierMath-Tiers-1-3-v2-Private. Task version: 2.0.0. Run ID: KUar3FaFjzq5SMKrEhtn8e. Started at 2026-06-17T02:30:14.000Z. Original task label: FrontierMath-Tiers-1-3-v2-Private. The published row reports mean_score=0.45263157894736844 and stderr=0.029536098269922092. The CSV has no sample-count column, so none is inferred. Standard error is not presented as a confidence interval. Meta-score selection: Per model, the highest documented reasoning tier is selected within the same Epoch source test design, original task, version, and scorer, independently of the measured value. This derived score copy does not claim matched compute budgets; the capture publishes neither a common budget nor all inference parameters. The raw model-profile observation remains unchanged. Selected reasoning tier: max.
Test family: GPQA Diamond
Reasoning · Family index: 46.2 / 100
Nominal family weight in the overall panel: 1.02 % · Coverage of this test family: 100 %
- Epoch AI benchmark run dataset snapshot
  Published configurations differ within this cohort.
  Measurement:
  epoch-gpqa-diamond
  Series weight within the test family:
  50 %
  Nominal contribution to the overall panel:
  0.51 %
  Raw value:
  0.8964646464646465 ratio
  Metric:
  choice
  Version:
  1.0.11
  Participating models:
  22
  Rank percentile within the cohort:
  21.4 %
  Retrieved:
  2026-10-19
  Comparison cohort:
  epoch-score-design-v1:["GPQA diamond","GPQA diamond","1.0.11","choice"]
  Reasoning tier:
  max
  Documented comparison setup
  Publication cohort:
  epoch-score-design-v1:["GPQA diamond","GPQA diamond","1.0.11","choice"]
  Dataset type:
  not-reported-in-captured-run-csv
  Setup and model parameters
  Published reasoning tier: max. The capture does not report other inference parameters or a compute budget.
  Independent evaluation. Epoch run alias: deepseek-v4-pro_max; Epoch model label: DeepSeek-V4-Pro; Display label field: DeepSeek v4 Pro (max); Reasoning setting encoded in the alias: max. Task ID: GPQA diamond. Task version: 1.0.11. Run ID: LuZUnhD5rBLjRw36DaqfvZ. Started at 2026-06-16T20:16:55.000Z. Original task label: GPQA diamond. The published row reports mean_score=0.8964646464646465 and stderr=0.017473014662966007. The CSV has no sample-count column, so none is inferred. Standard error is not presented as a confidence interval. Meta-score selection: Per model, the highest documented reasoning tier is selected within the same Epoch source test design, original task, version, and scorer, independently of the measured value. This derived score copy does not claim matched compute budgets; the capture publishes neither a common budget nor all inference parameters. The raw model-profile observation remains unchanged. Selected reasoning tier: max.
Test family: Humanity’s Last Exam
Reasoning · Family index: 68.3 / 100
Nominal family weight in the overall panel: 1.02 % · Coverage of this test family: 50 %
Test family: SWE-bench Verified
Coding · Family index: 53.4 / 100
Nominal family weight in the overall panel: 0.48 % · Coverage of this test family: 50 %
Test family: Terminal-Bench
Coding · Family index: 57.4 / 100
Nominal family weight in the overall panel: 0.48 % · Coverage of this test family: 100 %
Test family: Terminal-Bench Science
Coding · Family index: 34.8 / 100
Nominal family weight in the overall panel: 0.48 % · Coverage of this test family: 50 %
Test family: Code Migration
Coding · Family index: 45.2 / 100
Nominal family weight in the overall panel: 0.48 % · Coverage of this test family: 100 %
Test family: CorpFin v2
Finance · Family index: 52.7 / 100
Nominal family weight in the overall panel: 1.02 % · Coverage of this test family: 100 %
Test family: EMB
Finance · Family index: 30.5 / 100
Nominal family weight in the overall panel: 1.02 % · Coverage of this test family: 100 %
Test family: Finance Agent (v2)
Finance · Family index: 28.3 / 100
Nominal family weight in the overall panel: 1.02 % · Coverage of this test family: 100 %
Test family: LiveCodeBench v6
Coding · Family index: 85.7 / 100
Nominal family weight in the overall panel: 0.48 % · Coverage of this test family: 100 %
Test family: Harvey's Legal Agent Benchmark
Legal · Family index: 55.7 / 100
Nominal family weight in the overall panel: 1.79 % · Coverage of this test family: 100 %
Test family: LegalBench
Legal · Family index: 34.6 / 100
Nominal family weight in the overall panel: 1.79 % · Coverage of this test family: 100 %
Test family: Legal Research Bench
Legal · Family index: 31.7 / 100
Nominal family weight in the overall panel: 1.79 % · Coverage of this test family: 100 %
Test family: MedCode
Health · Family index: 29.9 / 100
Nominal family weight in the overall panel: 2.38 % · Coverage of this test family: 100 %
Test family: MedScribe
Health · Family index: 23.5 / 100
Nominal family weight in the overall panel: 2.38 % · Coverage of this test family: 100 %
Test family: MMLU-Pro
Knowledge · Family index: 71.4 / 100
Nominal family weight in the overall panel: 2.38 % · Coverage of this test family: 100 %
Test family: ProgramBench
Coding · Family index: 24.5 / 100
Nominal family weight in the overall panel: 0.48 % · Coverage of this test family: 100 %
Test family: ProofBench
Mathematics · Family index: 17.9 / 100
Nominal family weight in the overall panel: 1.43 % · Coverage of this test family: 100 %
Test family: SkillsBench
Agents · Family index: 44.8 / 100
Nominal family weight in the overall panel: 0.6 % · Coverage of this test family: 100 %
Test family: Public Benefits Bench v1.1
Productivity · Family index: 50.0 / 100
Nominal family weight in the overall panel: 3.57 % · Coverage of this test family: 100 %
Test family: Tax Agent Bench
Finance · Family index: 56.4 / 100
Nominal family weight in the overall panel: 1.02 % · Coverage of this test family: 100 %
Test family: TaxEval v2
Finance · Family index: 43.4 / 100
Nominal family weight in the overall panel: 1.02 % · Coverage of this test family: 100 %
Test family: Vibe Code Bench v1.1
Coding · Family index: 48.6 / 100
Nominal family weight in the overall panel: 0.48 % · Coverage of this test family: 100 %

Benchmark sources at a glance

Each evaluator measures different capabilities under its own conditions. Test series, composite indices and their components stay separate. The detailed results below retain their scales, configurations and source dates.

Evaluator	Test series	Composite indices	Recorded active measurements
Artificial Analysis	9	1	23
Vals AI	25	1	193
LMArena	2	0	66
Epoch AI	6	0	12

Coverage is not a quality ranking. Overlapping tests and multiple settings do not count as independent experiments. Vendor claims, historical and disputed results are excluded from this overview and remain labelled in the detailed evidence.

Published benchmarks

Finance

CorpFin v2

Sub-task	DeepSeek-V4-Pro (This model)	GPT-5.6 Luna	GPT-5.6 Sol
Exact Pages	59.21%	63.52% (best value in this row)	61.54%
Max Fitting Context	63.05%	64.1%	65.15% (best value in this row)
Overall	61.38%	64.22%	64.38% (best value in this row)
Shared Max Context	61.89%	65.04%	66.43% (best value in this row)

higher is betterAs of October 15, 2026Source: Vals AIbars scaled within each row, not across rows

Methodology and context

Metric: accuracy

Coding

LiveCodeBench v6

Sub-task	DeepSeek-V4-Pro (This model)	GPT-5.6 Sol
Easy	98.76% (best value in this row)	98.14%
Hard	72.57% (best value in this row)	67.43%
Medium	91.12% (best value in this row)	82.25%
Overall	87.48% (best value in this row)	82.6%

higher is betterAs of October 15, 2026Source: Vals AIbars scaled within each row, not across rows

Methodology and context

Metric: accuracy

Health

MedCode

Overall

1.GPT-6.1 Sol48.84%

2.GPT-6 Astra48.49%

3.GPT-6 Sol47.07%

4.GPT-6 Luna44.69%

5.DeepSeek-V4-Pro40.46%

higher is betterAs of October 15, 2026Scale: 0% - 100%

Source: Vals AI

|

gradually.ai

Methodology and context

Comparison cohort: vals-medcode:1:overall
Metric: accuracy
Scale: 0-100
Participants: 5
Methodology and context: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-10-13. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 382.667. Standard error: 2.122. Cost per task in USD: 0.06071. Source ID: vals-ai. Evidence capture ID: vals-medcode-2026-10-15-988a6ae1e07e. Source payload hash: 988a6ae1e07e33ce031572ed4a804be11c36ee97a1dd75cbe4740fff3b534f8c. Parser version: vals-astro-v3. Imported at: 2026-10-15

Source: Vals AI

MedScribe

Overall

1.GPT-6 Astra87.91%

2.GPT-6.1 Sol86.45%

3.GPT-6 Luna83.71%

4.GPT-6 Sol82.03%

5.DeepSeek-V4-Pro75.14%

higher is betterAs of October 15, 2026Scale: 0% - 100%

Source: Vals AI

|

gradually.ai

Methodology and context

Comparison cohort: vals-medscribe:1:overall
Metric: accuracy
Scale: 0-100
Participants: 5
Methodology and context: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-10-13. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 345.723. Standard error: 2.002. Cost per task in USD: 0.053954. Source ID: vals-ai. Evidence capture ID: vals-medscribe-2026-10-15-271fc122ddc2. Source payload hash: 271fc122ddc2b0602bd6c7947f65a9a7f0d8cab94f6dc27d62a9380a9317f9c0. Parser version: vals-astro-v3. Imported at: 2026-10-15

Source: Vals AI

ProofBench

Overall

1.GPT-6 Astra99%

1.GPT-6.1 Sol99%

3.GPT-5.6 Sol83%

5.GPT-5.6 Terra74%

6.GPT-6 Luna64%

7.GPT-5.6 Luna60%

8.DeepSeek-V4-Pro16%

higher is betterAs of October 15, 2026Scale: 0% - 100%

Source: Vals AI

|

gradually.ai

Methodology and context

Comparison cohort: vals-proof-bench:1.1:overall
Metric: accuracy
Scale: 0-100
Participants: 8
Methodology and context: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-10-12. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 423.45. Standard error: 3.685. Cost per task in USD: 0.033594. Source ID: vals-ai. Evidence capture ID: vals-proof_bench-2026-10-15-22be2aca25de. Source payload hash: 22be2aca25de5c4a9eb6448a95713e09a7c97126dfc6f4f75c4afd0e563209cb. Parser version: vals-astro-v3. Imported at: 2026-10-15

Source: Vals AI

Recorded measurement details

CorpFin v2: Exact Pages (Show measurement, test conditions, and source)

Source value: 59.207
Score: 59.207
Metric: accuracy
Unit: %
Benchmark version: 2
Category: finance
Task: exact_pages
Task label: Exact Pages
Direction: Higher is better
Comparison cohort: vals-corp-fin-v2:2:exact_pages
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-12. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 84.032. Standard error: 1.678. Cost per task in USD: 0.007781. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-10-15-7cf2217ab372. Source payload hash: 7cf2217ab3728d5a84bffce82aab9fc52d7f810b0cbe8c65b387df47f3d31c9b. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

CorpFin v2: Max Fitting Context (Show measurement, test conditions, and source)

Source value: 63.054
Score: 63.054
Metric: accuracy
Unit: %
Benchmark version: 2
Category: finance
Task: max_fitting_context
Task label: Max Fitting Context
Direction: Higher is better
Comparison cohort: vals-corp-fin-v2:2:max_fitting_context
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-12. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 79.933. Standard error: 1.648. Cost per task in USD: 0.316589. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-10-15-7cf2217ab372. Source payload hash: 7cf2217ab3728d5a84bffce82aab9fc52d7f810b0cbe8c65b387df47f3d31c9b. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

CorpFin v2: Overall (Show measurement, test conditions, and source)

Source value: 61.383
Score: 61.383
Metric: accuracy
Unit: %
Benchmark version: 2
Category: finance
Task: overall
Task label: Overall
Direction: Higher is better
Comparison cohort: vals-corp-fin-v2:2:overall
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-12. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 79.331. Standard error: 0.959. Cost per task in USD: 0.148855. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-10-15-7cf2217ab372. Source payload hash: 7cf2217ab3728d5a84bffce82aab9fc52d7f810b0cbe8c65b387df47f3d31c9b. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

CorpFin v2: Shared Max Context (Show measurement, test conditions, and source)

Source value: 61.888
Score: 61.888
Metric: accuracy
Unit: %
Benchmark version: 2
Category: finance
Task: shared_max_context
Task label: Shared Max Context
Direction: Higher is better
Comparison cohort: vals-corp-fin-v2:2:shared_max_context
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-08-12. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 74.027. Standard error: 1.658. Cost per task in USD: 0.122195. Source ID: vals-ai. Evidence capture ID: vals-corp_fin_v2-2026-10-15-7cf2217ab372. Source payload hash: 7cf2217ab3728d5a84bffce82aab9fc52d7f810b0cbe8c65b387df47f3d31c9b. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

LiveCodeBench v6: Easy (Show measurement, test conditions, and source)

Source value: 98.758
Score: 98.758
Metric: accuracy
Unit: %
Benchmark version: 1
Category: coding
Task: easy
Task label: Easy
Direction: Higher is better
Comparison cohort: vals-lcb:1:easy
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-09-01. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 73.491. Standard error: 0.617. Cost per task in USD: 0.010782. Source ID: vals-ai. Evidence capture ID: vals-lcb-2026-10-15-1a7c7dabc9d9. Source payload hash: 1a7c7dabc9d95a0182886b1edc42623987cac2095a0d407a5e334f37acdfa2f5. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

LiveCodeBench v6: Hard (Show measurement, test conditions, and source)

Source value: 72.571
Score: 72.571
Metric: accuracy
Unit: %
Benchmark version: 1
Category: coding
Task: hard
Task label: Hard
Direction: Higher is better
Comparison cohort: vals-lcb:1:hard
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-09-01. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 1487.685. Standard error: 2.385. Cost per task in USD: 0.23531. Source ID: vals-ai. Evidence capture ID: vals-lcb-2026-10-15-1a7c7dabc9d9. Source payload hash: 1a7c7dabc9d95a0182886b1edc42623987cac2095a0d407a5e334f37acdfa2f5. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

LiveCodeBench v6: Medium (Show measurement, test conditions, and source)

Source value: 91.123
Score: 91.123
Metric: accuracy
Unit: %
Benchmark version: 1
Category: coding
Task: medium
Task label: Medium
Direction: Higher is better
Comparison cohort: vals-lcb:1:medium
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-09-01. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 585.336. Standard error: 1.453. Cost per task in USD: 0.082915. Source ID: vals-ai. Evidence capture ID: vals-lcb-2026-10-15-1a7c7dabc9d9. Source payload hash: 1a7c7dabc9d95a0182886b1edc42623987cac2095a0d407a5e334f37acdfa2f5. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

LiveCodeBench v6: Overall (Show measurement, test conditions, and source)

Source value: 87.484
Score: 87.484
Metric: accuracy
Unit: %
Benchmark version: 1
Category: coding
Task: overall
Task label: Overall
Direction: Higher is better
Comparison cohort: vals-lcb:1:overall
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-09-01. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 715.504. Standard error: 0.953. Cost per task in USD: 0.109669. Source ID: vals-ai. Evidence capture ID: vals-lcb-2026-10-15-1a7c7dabc9d9. Source payload hash: 1a7c7dabc9d95a0182886b1edc42623987cac2095a0d407a5e334f37acdfa2f5. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

MedCode: Overall (Show measurement, test conditions, and source)

Source value: 40.455
Score: 40.455
Metric: accuracy
Unit: %
Benchmark version: 1
Category: health
Task: overall
Task label: Overall
Direction: Higher is better
Comparison cohort: vals-medcode:1:overall
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-10-13. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 382.667. Standard error: 2.122. Cost per task in USD: 0.06071. Source ID: vals-ai. Evidence capture ID: vals-medcode-2026-10-15-988a6ae1e07e. Source payload hash: 988a6ae1e07e33ce031572ed4a804be11c36ee97a1dd75cbe4740fff3b534f8c. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

MedScribe: Overall (Show measurement, test conditions, and source)

Source value: 75.144
Score: 75.144
Metric: accuracy
Unit: %
Benchmark version: 1
Category: health
Task: overall
Task label: Overall
Direction: Higher is better
Comparison cohort: vals-medscribe:1:overall
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-10-13. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 345.723. Standard error: 2.002. Cost per task in USD: 0.053954. Source ID: vals-ai. Evidence capture ID: vals-medscribe-2026-10-15-271fc122ddc2. Source payload hash: 271fc122ddc2b0602bd6c7947f65a9a7f0d8cab94f6dc27d62a9380a9317f9c0. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

ProofBench: Overall (Show measurement, test conditions, and source)

Source value: 16
Score: 16
Metric: accuracy
Unit: %
Benchmark version: 1.1
Category: math
Task: overall
Task label: Overall
Direction: Higher is better
Comparison cohort: vals-proof-bench:1.1:overall
Source type: independent-evaluator
Evaluator: Vals AI
Status: active
Retrieved at: 2026-10-15
Methodology: Source model ID: deepseek/deepseek-v4-pro. Source model name: DeepSeek-V4-Pro. Provider reported by source: DeepSeek. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-10-12. Reasoning effort: max. Maximum output tokens: 128000. Latency in seconds: 423.45. Standard error: 3.685. Cost per task in USD: 0.033594. Source ID: vals-ai. Evidence capture ID: vals-proof_bench-2026-10-15-22be2aca25de. Source payload hash: 22be2aca25de5c4a9eb6448a95713e09a7c97126dfc6f4f75c4afd0e563209cb. Parser version: vals-astro-v3. Imported at: 2026-10-15
Comparison methodology: Reasoning effort: max. Maximum output tokens: 128000

Page 1 of 30

Sources and data date

Type	Evidence and data date
Speed	DeepSeek, Reasoning, max effort. Output speed after the first streaming chunk, not total response time. API measurement of snapshot 0813. The unchanged API model name now serves this version, not the original April weights. Retrieved September 7, 2026. Source: Artificial Analysis · Speed (Artificial Analysis)Speed measurement methodology
Price	Off-peak API prices. Higher rates apply during the provider's peak hours. All rates and the UTC schedule are listed under Costs. Source: Pricing · API pricing (retrieved September 8, 2026)
Input	Documented modalities. Retrieved September 7, 2026. Source: Model documentation · Context window data · DeepSeek-V4-Pro model card · Editorial description reviewed October 2, 2026
Output	Documented modalities. Retrieved September 7, 2026. Source: Model documentation · Context window data · DeepSeek-V4-Pro model card · Editorial description reviewed October 2, 2026
Research status	Research date September 8, 2026. 4 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication.
Unresolved	release-date Original wording Exact launch date for the HF weights is not stated in the checked card; API alias is dated V4-Pro-0813.
Unresolved	base-model-benchmarks Original wording The 24 Base-model table observations are retained separately in /tmp/gradually-profile-research-deepseek-v4-base-benchmarks.json and intentionally excluded from this Instruct profile.
Additional source	Model metadata · DeepSeek V4 announcement · Editorial description reviewed October 2, 2026
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	Artificial Analysis (retrieved October 19, 2026)
Additional source	LMArena (retrieved October 15, 2026)
Additional source	LMArena (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	Vals AI (retrieved August 11, 2026; October 15, 2026)
Additional source	Vals AI (retrieved October 15, 2026)
Additional source	DeepSeek V4 model card (retrieved September 8, 2026)
Additional source	Epoch AI benchmark run dataset snapshot (retrieved October 19, 2026)
Additional source	DeepSeek V4.1 lifecycle notice · Editorial description reviewed October 2, 2026
Additional source	DeepSeek API changelog · Editorial description reviewed October 2, 2026
Additional source	OpenRouter: Popularity index (derived), CC BY 4.0
Additional source	Vercel AI Gateway: Popularity index (derived), CC BY 4.0
Additional source	Ollama: cumulative downloads for the popularity index
Additional source	Replicate: cumulative runs for the popularity index
Additional source	Google Trending Now: limited feed selection for Germany and the US
Additional source	Creative Commons Attribution 4.0