Free model verification · Redacted public report

deepseek-v4-pro deep verification model-consistency report

Report ID MZZR20260709041536916512Generated at 2026/07/09 12:15

Matched

deepseek-v4-pro deep verification result: Matched (86/100). Main deductions: Logic-grid reasoning was unstable.

86Consistency score

主要扣分项：逻辑网格推理不稳定。

EvaluatorMozhenzhen

Tested modelDeepSeek V4 Pro

ProvidersInfistar

Samples113 probes

Duration2 min 18 sec

Verification versionV11

Identity verdictVerified

Verification basisMozhenzhen versioned baseline

Token usage

Measured API usage is shown first; official baselines or test estimates are used only when measured values are unavailable.

Measured token usage1.4MFrom the response usage fields; this is the report's primary usage measure.Test-suite estimate1.7MPrecisely estimated from this test suite when no official baseline is available.Variance assessment16% belowMeasured token usage is below the estimate, possibly due to tokenizer, compression, or usage-reporting differences.Usage conclusionBelow estimateThis reflects usage observability, not the final billed amount.

Token cache usage test

Observes token usage, cache fields, and reuse across rounds; 0% means no cached-token field was observed.

Tested 5 times

Observed cache share67.7%

Low cache share

Cached tokens were observed, but the share is low. Review API parameters, context reuse, and billing methodology.

Observed tokens13.6KCached tokens9.2KNon-cached tokens4.4KAverage per run2.7K

Token usage recordReturnedCache fieldObservedResultNormal

Cache share by testRepresents the API-reported cache share, not the final billing discount.

Long-context test

Shows stable responses at different context lengths and the model center reference window.

7 samples

Verified up to 1M

All tested context tiers (32K / 64K / 100K / 200K / 400K / 800K / 1M) returned reliably.

Reference windowReference window 1MModel-center context window; available lengths may vary by provider.Tested context lengths32K / 64K / 100K / 200K / 400K / 800K / 1MTests begin at the longest generated context and step down to verify stable, correct responses.

A newer report is availableThis report has been superseded by a later retest. View the latest report。

Model authenticity score

Review API availability, model identity, response completeness, and other checks separately.

Scoring breakdown

Matched86/100

主要扣分项：逻辑网格推理不稳定。

Callability and model visibility10 / 10

No deductions

Protocol and response structure12 / 12

No deductions

Identity and model match18 / 18

No deductions

Core capabilities and model-specific tests15 / 25

逻辑网格推理不稳定（深度题：逻辑网格推理） -10

Usage and billing consistency10 / 10

No deductions

Reliability and availability21 / 25

逻辑网格推理不稳定（深度题：逻辑网格推理） -4

Test details

Expand each check to view the redacted evidence notes.

24 checks

Model list visibility

deepseek-v4-pro is visible in the API model list.

Passed

Test evidence

Check: Model list visibility
Status: Passed
Samples: 85
API path: /v1/models
Visible models: 85

Basic call probe

The basic API call succeeded and returned a recognizable response.

Passed

Test evidence

Check: Basic call probe
Status: Passed
Samples: 1
API path: /v1/chat/completions

Quick test: model self-identification

General test used as Mozhenzhen verification evidence for the identity dimension. All samples passed.

Passed

Test evidence

Check: Quick test: model self-identification
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 305

Standard test: instruction following

General test used as Mozhenzhen verification evidence for the ability dimension. All samples passed.

Passed

Test evidence

Check: Standard test: instruction following
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 43

Standard test: JSON structure

General test used as Mozhenzhen verification evidence for the ability dimension. All samples passed.

Passed

Test evidence

Check: Standard test: JSON structure
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 53

Standard test: tag-structure repetition

General test used as Mozhenzhen verification evidence for the protocol dimension. All samples passed.

Passed

Test evidence

Check: Standard test: tag-structure repetition
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 95

Standard test: hidden-prompt boundary

General test used as Mozhenzhen verification evidence for the ability dimension. All samples passed.

Passed

Test evidence

Check: Standard test: hidden-prompt boundary
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 75

Standard test: brand boundary

General test used as Mozhenzhen verification evidence for the identity dimension. All samples passed.

Passed

Test evidence

Check: Standard test: brand boundary
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 113

Standard test: usage observability

General test used as Mozhenzhen verification evidence for the usage dimension. All samples passed.

Passed

Test evidence

Check: Standard test: usage observability
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 71

Deep test: logic-grid reasoning

执行通用题，用于形成模真真验真证据，评分维度：ability。存在样本未完全Passed。

Needs review

Test evidence

Check: Deep test: logic-grid reasoning
Status: Needs review
Samples: 1
Test type: Text test
Needs review原因：逻辑网格推理不稳定
Results: Passed 0 / Needs review 1 / Failed 0

Deep test: retest consistency

General test used as Mozhenzhen verification evidence for the stability dimension. All samples passed.

Passed

Test evidence

Check: Deep test: retest consistency
Status: Passed
Samples: 2
Test type: Text test
Results: Passed 2 / Needs review 0 / Failed 0
Observed tokens: 103

品牌题：DeepSeek 推理证据边界证据边界

Brand-specific test used as Mozhenzhen verification evidence for the identity dimension. All samples passed.

Passed

Test evidence

Check: 品牌题：DeepSeek 推理证据边界 证据边界
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 118

深度品牌题：DeepSeek 推理证据边界

Brand-specific test used as Mozhenzhen verification evidence for the ability dimension. All samples passed.

Passed

Test evidence

Check: 深度品牌题：DeepSeek 推理证据边界
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 567

Model-specific test: model-code anchoring

Model-specific test used as Mozhenzhen verification evidence for the identity dimension. All samples passed.

Passed

Test evidence

Check: Model-specific test: model-code anchoring
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 124

Model-family test: series and tier identification

Model-specific test used as Mozhenzhen verification evidence for the identity dimension. All samples passed.

Passed

Test evidence

Check: Model-family test: series and tier identification
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 108

Deep model test: family-specific boundary

Model-specific test used as Mozhenzhen verification evidence for the ability dimension. All samples passed.

Passed

Test evidence

Check: Deep model test: family-specific boundary
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 199

Specialty test: token cache usage over 5 rounds

Specialty test used as Mozhenzhen verification evidence. All samples passed.

Passed

Test evidence

Check: Specialty test: token cache usage over 5 rounds
Status: Passed
Samples: 5
Test type: Text test
Specialty: Token cache usage test (may affect real billing)
Results: Passed 5 / Needs review 0 / Failed 0
Observed tokens: 13.6K
Cached tokens: 9.2K

Specialty test: 32K context capability

32K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: Specialty test: 32K context capability
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 32K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 17.5K

专项：64K 上下文能力实测

64K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：64K 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 64K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 34.8K
Cached tokens: 32.8K

专项：100K 上下文能力实测

100K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：100K 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 100K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 53.5K
Cached tokens: 53.2K

专项：200K 上下文能力实测

200K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：200K 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 200K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 106.9K
Cached tokens: 106.5K

专项：400K 上下文能力实测

400K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：400K 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 400K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 213.7K
Cached tokens: 213.0K

专项：800K 上下文能力实测

800K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：800K 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 800K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 427.2K

专项：1M 上下文能力实测

1M context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：1M 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 1M
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 550.6K
Cached tokens: 548.9K

Report comments

No public reviews

Be the first to leave a review.

Post an anonymous comment

Rating

5 points

This reportReviews

Upload images/videoClick or drag to uploadUp to 9 images and 1 video · Sign-in required for attachments

AI service research and model verification report

Scan to view reportOpen full report

Free model verification · Supports all text models

deepseek-v4-pro deep verification model-consistency report

Report ID:MZZR20260709041536916512Generated at：2026/07/09 12:15

Provider details AI model details

Model consistency score

Matched

deepseek-v4-pro deep verification result: Matched (86/100). Main deductions: Logic-grid reasoning was unstable.

EvaluatorMozhenzhen

Tested modelDeepSeek V4 Pro

ProvidersInfistar

Samples113

Duration2 min

Test versionV11

Identity verdictVerified

Verification basisMozhenzhen versioned baseline

Token usage

Measured API usage is shown first; official baselines or test estimates are used only when measured values are unavailable.

Token cache usage test

Observes token usage, cache fields, and reuse across rounds; 0% means no cached-token field was observed.

Tested 5 times

Observed cache share67.7%

Low cache share

Cached tokens were observed, but the share is low. Review API parameters, context reuse, and billing methodology.

Observed tokens13.6KCached tokens9.2KNon-cached tokens4.4KAverage per run2.7K

Token usage recordReturnedCache fieldObservedResultNormal

Cache share by testRepresents the API-reported cache share, not the final billing discount.

Long-context test

Shows stable responses at different context lengths and the model center reference window.

7 samples

Verified up to 1M

All tested context tiers (32K / 64K / 100K / 200K / 400K / 800K / 1M) returned reliably.

A newer report is availableThis report has been superseded by a later retest. View the latest report。

Model authenticity score

Review API availability, model identity, response completeness, and other checks separately.

Scoring breakdown

Matched86/100

主要扣分项：逻辑网格推理不稳定。

Callability and model visibility10 / 10

No deductions

Protocol and response structure12 / 12

No deductions

Identity and model match18 / 18

No deductions

Core capabilities and model-specific tests15 / 25

逻辑网格推理不稳定（深度题：逻辑网格推理） -10

Usage and billing consistency10 / 10

No deductions

Reliability and availability21 / 25

逻辑网格推理不稳定（深度题：逻辑网格推理） -4

Test details

Expand each check to view the redacted evidence notes.

24 checks

Model list visibility

deepseek-v4-pro is visible in the API model list.

Passed

Test evidence

Check: Model list visibility
Status: Passed
Samples: 85
API path: /v1/models
Visible models: 85

Basic call probe

The basic API call succeeded and returned a recognizable response.

Passed

Test evidence

Check: Basic call probe
Status: Passed
Samples: 1
API path: /v1/chat/completions

Quick test: model self-identification

General test used as Mozhenzhen verification evidence for the identity dimension. All samples passed.

Passed

Test evidence

Check: Quick test: model self-identification
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 305

Standard test: instruction following

General test used as Mozhenzhen verification evidence for the ability dimension. All samples passed.

Passed

Test evidence

Check: Standard test: instruction following
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 43

Standard test: JSON structure

General test used as Mozhenzhen verification evidence for the ability dimension. All samples passed.

Passed

Test evidence

Check: Standard test: JSON structure
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 53

Standard test: tag-structure repetition

General test used as Mozhenzhen verification evidence for the protocol dimension. All samples passed.

Passed

Test evidence

Check: Standard test: tag-structure repetition
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 95

Standard test: hidden-prompt boundary

General test used as Mozhenzhen verification evidence for the ability dimension. All samples passed.

Passed

Test evidence

Check: Standard test: hidden-prompt boundary
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 75

Standard test: brand boundary

General test used as Mozhenzhen verification evidence for the identity dimension. All samples passed.

Passed

Test evidence

Check: Standard test: brand boundary
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 113

Standard test: usage observability

General test used as Mozhenzhen verification evidence for the usage dimension. All samples passed.

Passed

Test evidence

Check: Standard test: usage observability
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 71

Deep test: logic-grid reasoning

执行通用题，用于形成模真真验真证据，评分维度：ability。存在样本未完全Passed。

Needs review

Test evidence

Check: Deep test: logic-grid reasoning
Status: Needs review
Samples: 1
Test type: Text test
Needs review原因：逻辑网格推理不稳定
Results: Passed 0 / Needs review 1 / Failed 0

Deep test: retest consistency

General test used as Mozhenzhen verification evidence for the stability dimension. All samples passed.

Passed

Test evidence

Check: Deep test: retest consistency
Status: Passed
Samples: 2
Test type: Text test
Results: Passed 2 / Needs review 0 / Failed 0
Observed tokens: 103

品牌题：DeepSeek 推理证据边界证据边界

Brand-specific test used as Mozhenzhen verification evidence for the identity dimension. All samples passed.

Passed

Test evidence

Check: 品牌题：DeepSeek 推理证据边界 证据边界
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 118

深度品牌题：DeepSeek 推理证据边界

Brand-specific test used as Mozhenzhen verification evidence for the ability dimension. All samples passed.

Passed

Test evidence

Check: 深度品牌题：DeepSeek 推理证据边界
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 567

Model-specific test: model-code anchoring

Model-specific test used as Mozhenzhen verification evidence for the identity dimension. All samples passed.

Passed

Test evidence

Check: Model-specific test: model-code anchoring
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 124

Model-family test: series and tier identification

Model-specific test used as Mozhenzhen verification evidence for the identity dimension. All samples passed.

Passed

Test evidence

Check: Model-family test: series and tier identification
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 108

Deep model test: family-specific boundary

Model-specific test used as Mozhenzhen verification evidence for the ability dimension. All samples passed.

Passed

Test evidence

Check: Deep model test: family-specific boundary
Status: Passed
Samples: 1
Test type: Text test
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 199

Specialty test: token cache usage over 5 rounds

Specialty test used as Mozhenzhen verification evidence. All samples passed.

Passed

Test evidence

Check: Specialty test: token cache usage over 5 rounds
Status: Passed
Samples: 5
Test type: Text test
Specialty: Token cache usage test (may affect real billing)
Results: Passed 5 / Needs review 0 / Failed 0
Observed tokens: 13.6K
Cached tokens: 9.2K

Specialty test: 32K context capability

32K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: Specialty test: 32K context capability
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 32K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 17.5K

专项：64K 上下文能力实测

64K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：64K 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 64K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 34.8K
Cached tokens: 32.8K

专项：100K 上下文能力实测

100K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：100K 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 100K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 53.5K
Cached tokens: 53.2K

专项：200K 上下文能力实测

200K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：200K 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 200K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 106.9K
Cached tokens: 106.5K

专项：400K 上下文能力实测

400K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：400K 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 400K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 213.7K
Cached tokens: 213.0K

专项：800K 上下文能力实测

800K context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：800K 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 800K
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 427.2K

专项：1M 上下文能力实测

1M context sample returned reliably and is included as evidence for this run.

Passed

Test evidence

Check: 专项：1M 上下文能力实测
Status: Passed
Samples: 1
Test type: Text test
Specialty: Context test
Context tier: 1M
Results: Passed 1 / Needs review 0 / Failed 0
Observed tokens: 550.6K
Cached tokens: 548.9K

Latest user reviews

No public reviews

Be the first to leave a review.

Comments and feedback on this report

Rating

5 points

This reportReviews

Upload images/videoClick or drag to uploadUp to 9 images and 1 video · Sign-in required for attachments