claude-opus-4-8 deep verification model-consistency report
Report ID MZZR20260718024248008250Generated at 2026/07/18 10:42
Matched with issues
claude-opus-4-8 deep verification result: Matched with issues (80/100). Main deductions: Model-code anchoring was unstable; Model-family identification was unstable.
80Consistency score
主要扣分项：模型代码锚定不稳定；模型族识别不稳定。
EvaluatorMozhenzhen
Tested modelClaude Opus 4.8
ProvidersInfistar
Samples33 probes
Duration3 min 28 sec
Verification versionV10
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 usage2.4MFrom the response usage fields; this is the report's primary usage measure.Test-suite estimate74.9KPrecisely estimated from this test suite when no official baseline is available.Variance assessment3096% aboveMeasured token usage is above the estimate; review billing or provider logs.Usage conclusionAbove 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.
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 950K
All tested context tiers (950K / 800K / 400K / 200K / 100K / 64K / 32K) returned reliably.
Reference windowReference window 1MModel-center context window; available lengths may vary by provider.Tested context lengths950K / 800K / 400K / 200K / 100K / 64K / 32KTests 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A newer report is availableThis report has been superseded by a later retest.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.
claude-opus-4-8 deep verification result: Matched with issues (80/100). Main deductions: Model-code anchoring was unstable; Model-family identification was unstable.
EvaluatorMozhenzhen
Tested modelClaude Opus 4.8
ProvidersInfistar
Samples33
Duration3 min
Test versionV10
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 usage2.4MFrom the response usage fields; this is the report's primary usage measure.Test-suite estimate74.9KPrecisely estimated from this test suite when no official baseline is available.Variance assessment3096% aboveMeasured token usage is above the estimate; review billing or provider logs.Usage conclusionAbove 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.
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 950K
All tested context tiers (950K / 800K / 400K / 200K / 100K / 64K / 32K) returned reliably.
Reference windowReference window 1MModel-center context window; available lengths may vary by provider.Tested context lengths950K / 800K / 400K / 200K / 100K / 64K / 32KTests 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.
