ModelVerdict

Comparison

Gemma 4 31B vs Mistral Medium 3.5

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 extraction97.6% / 97.8%96.1% / 93.5%94.9% / 97.4%95.4% / 97.0%93.6% / 98.2%
Email triage96.5% / 94.9%97.1% / 94.5%95.1% / 92.7%——
Personal data detection99.9% / 99.7%99.8% / 99.8%99.9% / 100.0%——
Contract clauses99.6% / 98.1%100.0% / 100.0%94.4% / 97.1%——

Correct fields: Gemma 4 31B / Mistral Medium 3.5; the better one in bold. Click a cell to see the details.

Invoice extraction · English (US) (150 documents)

Mistral Medium 3.5 is more accurate by 0.1 pp. Gemma 4 31B is 26.1× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields97.6%97.8%
Error-free documents81%79%
Per 1,000 documents$0.17$4.37
To fix / 1,000187207
Speed (median)8.9 sec2.3 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Supplier99.3%100.0%
Supplier ID (EIN, IČO…)98.0%100.0%
Invoice number100.0%100.0%
Payment reference99.3%94.7%
Invoice date98.7%98.7%
Due date98.7%99.3%
Net by tax rate95.3%88.0%
Tax by rate98.7%98.7%
Total100.0%100.0%
Currency100.0%100.0%
Bank account86.0%96.0%

Documents where they differ

Invoice extraction · Czech (150 documents)

Gemma 4 31B is more accurate by 2.6 pp. Gemma 4 31B is 22.1× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields96.1%93.5%
Error-free documents67%63%
Per 1,000 documents$0.21$4.69
To fix / 1,000333367
Speed (median)10.4 sec2.6 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Supplier96.7%90.0%
Supplier ID (EIN, IČO…)93.3%87.3%
VAT ID96.7%85.3%
Invoice number100.0%99.3%
Payment reference96.0%95.3%
Invoice date96.0%88.7%
Tax point95.3%86.7%
Due date94.0%90.7%
Net by tax rate98.0%99.3%
Tax by rate97.3%98.7%
Total99.3%100.0%
Currency100.0%100.0%
Bank account86.7%94.0%

Documents where they differ

Invoice extraction · German (150 documents)

Mistral Medium 3.5 is more accurate by 2.5 pp. Gemma 4 31B is 23.7× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields94.9%97.4%
Error-free documents64%75%
Per 1,000 documents$0.20$4.72
To fix / 1,000360253
Speed (median)7.7 sec2.5 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Supplier98.7%97.3%
Supplier ID (EIN, IČO…)99.3%98.7%
VAT ID94.0%96.0%
Invoice number99.3%100.0%
Payment reference98.7%92.7%
Invoice date87.3%98.0%
Tax point92.0%90.7%
Due date90.7%97.3%
Net by tax rate94.7%98.0%
Tax by rate97.3%100.0%
Total99.3%100.0%
Currency99.3%100.0%
Bank account82.7%97.3%

Documents where they differ

Invoice extraction · Spanish (87 documents)

Mistral Medium 3.5 is more accurate by 1.6 pp. Gemma 4 31B is 23.3× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields95.4%97.0%
Error-free documents54%67%
Per 1,000 documents$0.20$4.75
To fix / 1,000460330
Speed (median)8.7 sec2.5 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Supplier95.4%99.0%
Supplier ID (EIN, IČO…)93.1%100.0%
VAT ID98.9%94.0%
Invoice number100.0%100.0%
Payment reference95.4%93.0%
Invoice date98.9%100.0%
Tax point100.0%93.0%
Due date97.7%97.0%
Net by tax rate81.6%85.0%
Tax by rate100.0%100.0%
Total100.0%100.0%
Currency100.0%100.0%
Bank account79.3%100.0%

Documents where they differ

Invoice extraction · French (87 documents)

Mistral Medium 3.5 is more accurate by 4.6 pp. Gemma 4 31B is 21.7× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields93.6%98.2%
Error-free documents51%81%
Per 1,000 documents$0.22$4.81
To fix / 1,000494190
Speed (median)7.2 sec2.6 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Supplier96.6%99.0%
Supplier ID (EIN, IČO…)77.0%99.0%
VAT ID95.4%98.0%
Invoice number97.7%100.0%
Payment reference96.6%94.0%
Invoice date96.6%100.0%
Tax point94.3%100.0%
Due date96.6%100.0%
Net by tax rate93.1%95.0%
Tax by rate100.0%98.0%
Total100.0%99.0%
Currency98.9%96.0%
Bank account73.6%98.0%

Documents where they differ

Email triage · English (US) (150 documents)

Gemma 4 31B is more accurate by 1.6 pp. Gemma 4 31B is 12.2× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields96.5%94.9%
Error-free documents72%57%
Per 1,000 documents$0.12$1.47
To fix / 1,000280427
Speed (median)2.2 sec0.8 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Category100.0%100.0%
Priority85.3%63.3%
Sentiment83.3%90.7%
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

  • Cody Baldwin · Email · Gemma 4 31B 9/9, Mistral Medium 3.5 7/9
  • Allison Floyd · Email · Gemma 4 31B 7/9, Mistral Medium 3.5 9/9
  • Danielle Little · Email · Gemma 4 31B 9/9, Mistral Medium 3.5 7/9
  • Russell Bates · Email · Gemma 4 31B 9/9, Mistral Medium 3.5 8/9
  • Kyle Schwartz · Email · Gemma 4 31B 9/9, Mistral Medium 3.5 8/9
  • James Arnold · Email · Gemma 4 31B 9/9, Mistral Medium 3.5 8/9
  • Sheila Evans · Email · Gemma 4 31B 8/9, Mistral Medium 3.5 9/9
  • Jason Davis · Email · Gemma 4 31B 8/9, Mistral Medium 3.5 9/9

Email triage · Czech (150 documents)

Gemma 4 31B is more accurate by 2.6 pp. Gemma 4 31B is 12.3× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields97.1%94.5%
Error-free documents75%55%
Per 1,000 documents$0.16$1.95
To fix / 1,000253447
Speed (median)2.5 sec0.8 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Category100.0%100.0%
Priority85.3%63.3%
Sentiment88.7%87.3%
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)

Gemma 4 31B is more accurate by 2.4 pp. Gemma 4 31B is 11.9× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields95.1%92.7%
Error-free documents62%48%
Per 1,000 documents$0.15$1.77
To fix / 1,000380520
Speed (median)3.6 sec0.8 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Category100.0%100.0%
Priority78.0%56.7%
Sentiment78.0%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

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

Gemma 4 31B is more accurate by 0.2 pp. Gemma 4 31B is 13.6× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields99.9%99.7%
Error-free documents99%98%
Per 1,000 documents$0.08$1.15
To fix / 1,000720
Speed (median)2.4 sec0.8 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Names100.0%99.3%
Email addresses100.0%100.0%
Phone numbers100.0%100.0%
Addresses99.3%98.7%
Dates of birth100.0%100.0%
National IDs100.0%100.0%
Bank accounts100.0%100.0%

Documents where they differ

Personal data detection · Czech (150 documents)

Both are equally accurate. Gemma 4 31B is 11.8× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields99.8%99.8%
Error-free documents99%99%
Per 1,000 documents$0.12$1.40
To fix / 1,0001313
Speed (median)2.9 sec0.8 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Names100.0%98.7%
Email addresses100.0%100.0%
Phone numbers100.0%100.0%
Addresses98.7%100.0%
Dates of birth100.0%100.0%
National IDs100.0%100.0%
Bank accounts100.0%100.0%

Documents where they differ

Personal data detection · German (150 documents)

Mistral Medium 3.5 is more accurate by 0.1 pp. Gemma 4 31B is 13.6× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields99.9%100.0%
Error-free documents99%100%
Per 1,000 documents$0.10$1.38
To fix / 1,00070
Speed (median)1.9 sec0.9 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Names100.0%100.0%
Email addresses100.0%100.0%
Phone numbers100.0%100.0%
Addresses99.3%100.0%
Dates of birth100.0%100.0%
National IDs100.0%100.0%
Bank accounts100.0%100.0%

Documents where they differ

Contract clauses · English (US) (150 documents)

Gemma 4 31B is more accurate by 1.5 pp. Gemma 4 31B is 13× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields99.6%98.1%
Error-free documents97%87%
Per 1,000 documents$0.18$2.35
To fix / 1,00027127
Speed (median)2.5 sec1.1 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Provider98.0%87.3%
Provider ID100.0%100.0%
Client98.0%88.7%
Client ID100.0%99.3%
Contract type100.0%100.0%
Price (net)100.0%100.0%
Currency100.0%100.0%
Billing period100.0%100.0%
Effective date99.3%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

Contract clauses · Czech (150 documents)

Both are equally accurate. Gemma 4 31B is 12× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields100.0%100.0%
Error-free documents99%100%
Per 1,000 documents$0.25$3.06
To fix / 1,00070
Speed (median)6.0 sec1.2 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
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 date99.3%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

Contract clauses · German (150 documents)

Mistral Medium 3.5 is more accurate by 2.7 pp. Gemma 4 31B is 12.5× cheaper.

MetricGemma 4 31BMistral Medium 3.5
Correct fields94.4%97.1%
Error-free documents27%75%
Per 1,000 documents$0.22$2.76
To fix / 1,000727247
Speed (median)2.6 sec1.2 sec
Valid answers100.0%100.0%

Accuracy per field

FieldGemma 4 31BMistral Medium 3.5
Provider100.0%100.0%
Provider ID100.0%82.0%
Client100.0%100.0%
Client ID100.0%80.7%
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)27.3%100.0%
Governing law100.0%100.0%

Documents where they differ