ModelVerdict

Comparison

Claude Sonnet 5 vs DeepSeek V4.1 Flash

Same documents, same prompt: where each model is right, what it really costs and how fast it answers.

At a glance

Task · documentsEnglish (US)CzechGermanSpanishFrench
Invoice extraction99.8% / 99.4%100.0% / 99.9%99.0% / 99.5%93.6% / 98.3%98.7% / 99.4%
Email triage98.2% / 97.6%98.2% / 97.4%97.9% / 96.3%——
Personal data detection99.7% / 99.6%98.9% / 99.7%99.7% / 99.8%——
Contract clauses100.0% / 100.0%100.0% / 100.0%98.2% / 97.2%——

Correct fields: Claude Sonnet 5 / DeepSeek V4.1 Flash; the better one in bold. Click a cell to see the details.

Invoice extraction · English (US) (150 documents)

Claude Sonnet 5 is more accurate by 0.4 pp. DeepSeek V4.1 Flash is 17× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields99.8%99.4%
Error-free documents99%93%
Per 1,000 documents$8.84$0.52
To fix / 1,0001367
Speed (median)3.2 sec5.7 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Supplier100.0%100.0%
Supplier ID (EIN, IČO…)100.0%100.0%
Invoice number100.0%100.0%
Payment reference100.0%96.7%
Invoice date100.0%100.0%
Due date100.0%100.0%
Net by tax rate98.7%100.0%
Tax by rate99.3%100.0%
Total100.0%100.0%
Currency100.0%99.3%
Bank account100.0%97.3%

Documents where they differ

Invoice extraction · Czech (150 documents)

Claude Sonnet 5 is more accurate by 0.1 pp. DeepSeek V4.1 Flash is 19× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields100.0%99.9%
Error-free documents100%99%
Per 1,000 documents$9.41$0.50
To fix / 1,000013
Speed (median)3.6 sec5.0 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Supplier100.0%99.3%
Supplier ID (EIN, IČO…)100.0%100.0%
VAT ID100.0%100.0%
Invoice number100.0%100.0%
Payment reference100.0%100.0%
Invoice date100.0%100.0%
Tax point100.0%100.0%
Due date100.0%100.0%
Net by tax rate100.0%100.0%
Tax by rate100.0%100.0%
Total100.0%100.0%
Currency100.0%100.0%
Bank account100.0%99.3%

Documents where they differ

Invoice extraction · German (150 documents)

DeepSeek V4.1 Flash is more accurate by 0.6 pp. DeepSeek V4.1 Flash is 18.4× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields99.0%99.5%
Error-free documents87%94%
Per 1,000 documents$10.42$0.57
To fix / 1,00013360
Speed (median)3.8 sec6.4 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Supplier100.0%99.3%
Supplier ID (EIN, IČO…)100.0%98.7%
VAT ID100.0%99.3%
Invoice number100.0%100.0%
Payment reference100.0%100.0%
Invoice date99.3%100.0%
Tax point100.0%99.3%
Due date100.0%99.3%
Net by tax rate89.3%100.0%
Tax by rate100.0%100.0%
Total100.0%100.0%
Currency100.0%100.0%
Bank account98.0%98.0%

Documents where they differ

Invoice extraction · Spanish (100 documents)

DeepSeek V4.1 Flash is more accurate by 4.6 pp. DeepSeek V4.1 Flash is 16.8× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields93.6%98.3%
Error-free documents77%77%
Per 1,000 documents$11.89$0.71
To fix / 1,000230228
Speed (median)4.0 sec5.8 sec
Valid answers95.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Supplier94.0%93.7%
Supplier ID (EIN, IČO…)95.0%100.0%
VAT ID95.0%100.0%
Invoice number95.0%100.0%
Payment reference94.0%97.5%
Invoice date95.0%100.0%
Tax point95.0%98.7%
Due date95.0%100.0%
Net by tax rate79.0%88.6%
Tax by rate95.0%100.0%
Total95.0%100.0%
Currency95.0%100.0%
Bank account95.0%98.7%

Documents where they differ

Invoice extraction · French (97 documents)

DeepSeek V4.1 Flash is more accurate by 0.7 pp. DeepSeek V4.1 Flash is 17× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields98.7%99.4%
Error-free documents96%92%
Per 1,000 documents$11.02$0.65
To fix / 1,0004181
Speed (median)4.4 sec5.7 sec
Valid answers99.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Supplier99.0%100.0%
Supplier ID (EIN, IČO…)99.0%100.0%
VAT ID99.0%98.7%
Invoice number99.0%100.0%
Payment reference97.9%98.7%
Invoice date99.0%100.0%
Tax point99.0%100.0%
Due date99.0%100.0%
Net by tax rate96.9%100.0%
Tax by rate99.0%100.0%
Total99.0%100.0%
Currency99.0%100.0%
Bank account97.9%94.6%

Documents where they differ

Email triage · English (US) (150 documents)

Claude Sonnet 5 is more accurate by 0.6 pp. DeepSeek V4.1 Flash is 17.2× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields98.2%97.6%
Error-free documents84%79%
Per 1,000 documents$4.01$0.23
To fix / 1,000160213
Speed (median)3.8 sec3.8 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Category100.0%100.0%
Priority97.3%96.7%
Sentiment86.7%82.0%
Customer name100.0%100.0%
Order number100.0%100.0%
Invoice number100.0%100.0%
Amount100.0%100.0%
Currency100.0%100.0%
Deadline100.0%100.0%

Documents where they differ

  • Russell Bates · Email · Claude Sonnet 5 9/9, DeepSeek V4.1 Flash 8/9
  • Allison Floyd · Email · Claude Sonnet 5 8/9, DeepSeek V4.1 Flash 9/9
  • Robert Lewis · Email · Claude Sonnet 5 9/9, DeepSeek V4.1 Flash 8/9
  • Angela Harmon · Email · Claude Sonnet 5 9/9, DeepSeek V4.1 Flash 8/9

Email triage · Czech (150 documents)

Claude Sonnet 5 is more accurate by 0.7 pp. DeepSeek V4.1 Flash is 16.6× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields98.2%97.4%
Error-free documents83%77%
Per 1,000 documents$4.91$0.30
To fix / 1,000167227
Speed (median)4.7 sec4.2 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Category100.0%100.0%
Priority96.0%98.7%
Sentiment87.3%78.0%
Customer name100.0%100.0%
Order number100.0%100.0%
Invoice number100.0%100.0%
Amount100.0%100.0%
Currency100.0%100.0%
Deadline100.0%100.0%

Documents where they differ

Email triage · German (150 documents)

Claude Sonnet 5 is more accurate by 1.5 pp. DeepSeek V4.1 Flash is 16.9× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields97.9%96.3%
Error-free documents81%68%
Per 1,000 documents$5.11$0.30
To fix / 1,000187324
Speed (median)4.3 sec4.2 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Category100.0%100.0%
Priority96.0%97.2%
Sentiment84.7%69.7%
Customer name100.0%100.0%
Order number100.0%100.0%
Invoice number100.0%100.0%
Amount100.0%100.0%
Currency100.0%100.0%
Deadline100.0%100.0%

Documents where they differ

Personal data detection · English (US) (150 documents)

Claude Sonnet 5 is more accurate by 0.1 pp. DeepSeek V4.1 Flash is 12.3× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields99.7%99.6%
Error-free documents98%97%
Per 1,000 documents$3.50$0.29
To fix / 1,0002027
Speed (median)2.5 sec3.8 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Names100.0%98.7%
Email addresses100.0%100.0%
Phone numbers100.0%100.0%
Addresses98.0%98.7%
Dates of birth100.0%100.0%
National IDs100.0%100.0%
Bank accounts100.0%100.0%

Documents where they differ

  • Margaret Mills · Text · Claude Sonnet 5 7/7, DeepSeek V4.1 Flash 6/7

Personal data detection · Czech (150 documents)

DeepSeek V4.1 Flash is more accurate by 0.8 pp. DeepSeek V4.1 Flash is 12.9× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields98.9%99.7%
Error-free documents92%98%
Per 1,000 documents$4.06$0.31
To fix / 1,0008020
Speed (median)3.2 sec4.2 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Names96.7%100.0%
Email addresses100.0%100.0%
Phone numbers100.0%100.0%
Addresses97.3%98.0%
Dates of birth98.0%100.0%
National IDs100.0%100.0%
Bank accounts100.0%100.0%

Documents where they differ

Personal data detection · German (150 documents)

DeepSeek V4.1 Flash is more accurate by 0.1 pp. DeepSeek V4.1 Flash is 14.4× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields99.7%99.8%
Error-free documents98%99%
Per 1,000 documents$4.40$0.31
To fix / 1,0002013
Speed (median)3.7 sec3.7 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Names98.7%100.0%
Email addresses100.0%100.0%
Phone numbers100.0%100.0%
Addresses100.0%98.7%
Dates of birth99.3%100.0%
National IDs100.0%100.0%
Bank accounts100.0%100.0%

Documents where they differ

  • Gunnar Zorbach · Text · Claude Sonnet 5 7/7, DeepSeek V4.1 Flash 6/7

Contract clauses · English (US) (150 documents)

Both are equally accurate. DeepSeek V4.1 Flash is 15.7× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields100.0%100.0%
Error-free documents100%100%
Per 1,000 documents$5.04$0.32
To fix / 1,00000
Speed (median)3.4 sec3.9 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Provider100.0%100.0%
Provider ID100.0%100.0%
Client100.0%100.0%
Client ID100.0%100.0%
Contract type100.0%100.0%
Price (net)100.0%100.0%
Currency100.0%100.0%
Billing period100.0%100.0%
Effective date100.0%100.0%
End date100.0%100.0%
Notice period (months)100.0%100.0%
Late-payment penalty (% per day)100.0%100.0%
Governing law100.0%100.0%

Documents where they differ

Both models got the same number of fields right on every example document.

Contract clauses · Czech (150 documents)

Both are equally accurate. DeepSeek V4.1 Flash is 14.4× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields100.0%100.0%
Error-free documents100%100%
Per 1,000 documents$6.05$0.42
To fix / 1,00000
Speed (median)3.6 sec4.3 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Provider100.0%100.0%
Provider ID100.0%100.0%
Client100.0%100.0%
Client ID100.0%100.0%
Contract type100.0%100.0%
Price (net)100.0%100.0%
Currency100.0%100.0%
Billing period100.0%100.0%
Effective date100.0%100.0%
End date100.0%100.0%
Notice period (months)100.0%100.0%
Late-payment penalty (% per day)100.0%100.0%
Governing law100.0%100.0%

Documents where they differ

Both models got the same number of fields right on every example document.

Contract clauses · German (150 documents)

Claude Sonnet 5 is more accurate by 1 pp. DeepSeek V4.1 Flash is 13.9× cheaper.

MetricClaude Sonnet 5DeepSeek V4.1 Flash
Correct fields98.2%97.2%
Error-free documents87%77%
Per 1,000 documents$7.11$0.51
To fix / 1,000133228
Speed (median)4.3 sec4.6 sec
Valid answers100.0%100.0%

Accuracy per field

FieldClaude Sonnet 5DeepSeek V4.1 Flash
Provider100.0%100.0%
Provider ID87.3%81.4%
Client100.0%100.0%
Client ID89.3%82.8%
Contract type100.0%100.0%
Price (net)100.0%100.0%
Currency100.0%100.0%
Billing period100.0%100.0%
Effective date100.0%100.0%
End date100.0%100.0%
Notice period (months)100.0%100.0%
Late-payment penalty (% per day)100.0%100.0%
Governing law100.0%100.0%

Documents where they differ