Claude Sonnet 5 vs GPT-6 Luna
Same documents, same prompt: where each model is right, what it really costs and how fast it answers.
At a glance
| Task · documents | English (US) | Czech | German | Spanish | French |
|---|---|---|---|---|---|
| Invoice extraction | 99.8% / 99.7% | 100.0% / 99.3% | 99.0% / 99.2% | 93.6% / 99.1% | 98.7% / 98.9% |
| Email triage | 98.2% / 97.5% | 98.2% / 97.9% | 97.9% / 97.0% | — | — |
| Personal data detection | 99.7% / 99.6% | 98.9% / 99.6% | 99.7% / 99.6% | — | — |
| Contract clauses | 100.0% / 99.8% | 100.0% / 100.0% | 98.2% / 100.0% | — | — |
Correct fields: Claude Sonnet 5 / GPT-6 Luna; 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 0.1 pp. GPT-6 Luna is 17.4× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.8% | 99.7% |
| Error-free documents | 99% | 97% |
| Per 1,000 documents | $8.84 | $0.51 |
| To fix / 1,000 | 13 | 33 |
| Speed (median) | 3.2 sec | 3.4 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Supplier | 100.0% | 100.0% |
| Supplier ID (EIN, IČO…) | 100.0% | 100.0% |
| Invoice number | 100.0% | 99.3% |
| Payment reference | 100.0% | 98.0% |
| Invoice date | 100.0% | 100.0% |
| Due date | 100.0% | 100.0% |
| Net by tax rate | 98.7% | 100.0% |
| Tax by rate | 99.3% | 100.0% |
| Total | 100.0% | 100.0% |
| Currency | 100.0% | 100.0% |
| Bank account | 100.0% | 99.3% |
Documents where they differ
- Raleigh Engineering Inc. · INV253511 · Scan · Claude Sonnet 5 11/11, GPT-6 Luna 10/11
Invoice extraction · Czech (150 documents)
Claude Sonnet 5 is more accurate by 0.7 pp. GPT-6 Luna is 18.6× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.0% | 99.3% |
| Error-free documents | 100% | 93% |
| Per 1,000 documents | $9.41 | $0.51 |
| To fix / 1,000 | 0 | 67 |
| Speed (median) | 3.6 sec | 3.2 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Supplier | 100.0% | 95.3% |
| Supplier ID (EIN, IČO…) | 100.0% | 98.7% |
| VAT ID | 100.0% | 98.0% |
| Invoice number | 100.0% | 100.0% |
| Payment reference | 100.0% | 100.0% |
| Invoice date | 100.0% | 100.0% |
| Tax point | 100.0% | 100.0% |
| Due date | 100.0% | 99.3% |
| Net by tax rate | 100.0% | 100.0% |
| Tax by rate | 100.0% | 100.0% |
| Total | 100.0% | 100.0% |
| Currency | 100.0% | 100.0% |
| Bank account | 100.0% | 99.3% |
Documents where they differ
- Tesař s.r.o. · 20250237 · Scan · Claude Sonnet 5 13/13, GPT-6 Luna 10/13
- Holub s.r.o. · FV2500655 · Scan · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Slavěna Pospíšilová · 26200409 · Scan · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Agrotrade s.r.o. · DB-2025-00296 · Clean image · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Stavebniny Sýkora s.r.o. · FV-2025-00537 · Scan · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Východočeská obchodní s.r.o. · FV2501176 · Photo · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
Invoice extraction · German (150 documents)
GPT-6 Luna is more accurate by 0.3 pp. GPT-6 Luna is 19.1× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.0% | 99.2% |
| Error-free documents | 87% | 91% |
| Per 1,000 documents | $10.42 | $0.54 |
| To fix / 1,000 | 133 | 87 |
| Speed (median) | 3.8 sec | 3.7 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Supplier | 100.0% | 100.0% |
| Supplier ID (EIN, IČO…) | 100.0% | 98.7% |
| VAT ID | 100.0% | 99.3% |
| Invoice number | 100.0% | 99.3% |
| Payment reference | 100.0% | 100.0% |
| Invoice date | 99.3% | 98.7% |
| Tax point | 100.0% | 100.0% |
| Due date | 100.0% | 100.0% |
| Net by tax rate | 89.3% | 100.0% |
| Tax by rate | 100.0% | 100.0% |
| Total | 100.0% | 100.0% |
| Currency | 100.0% | 100.0% |
| Bank account | 98.0% | 94.0% |
Documents where they differ
- Druckerei Baumann GmbH · 2025-0324 · Scan · Claude Sonnet 5 12/13, GPT-6 Luna 13/13
- Zimmermann & Schulz GmbH · RE11274 · Scan · Claude Sonnet 5 12/13, GPT-6 Luna 13/13
- Malerbetrieb Otfried Haas · RE-2025-00168 · Scan · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Braun IT-Service GmbH · RE2500825 · Photo · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Südform AG · 2025-0307 · PDF text · Claude Sonnet 5 12/13, GPT-6 Luna 13/13
- Wesermetall GmbH · RE2501375 · Scan · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Zimmermann UG (haftungsbeschränkt) · GS10958 · Clean image · Claude Sonnet 5 12/13, GPT-6 Luna 13/13
Invoice extraction · Spanish (100 documents)
GPT-6 Luna is more accurate by 5.5 pp. GPT-6 Luna is 21.5× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 93.6% | 99.1% |
| Error-free documents | 77% | 88% |
| Per 1,000 documents | $11.89 | $0.55 |
| To fix / 1,000 | 230 | 120 |
| Speed (median) | 4.0 sec | 3.9 sec |
| Valid answers | 95.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Supplier | 94.0% | 99.0% |
| Supplier ID (EIN, IČO…) | 95.0% | 100.0% |
| VAT ID | 95.0% | 100.0% |
| Invoice number | 95.0% | 100.0% |
| Payment reference | 94.0% | 98.0% |
| Invoice date | 95.0% | 100.0% |
| Tax point | 95.0% | 100.0% |
| Due date | 95.0% | 100.0% |
| Net by tax rate | 79.0% | 100.0% |
| Tax by rate | 95.0% | 100.0% |
| Total | 95.0% | 100.0% |
| Currency | 95.0% | 100.0% |
| Bank account | 95.0% | 91.0% |
Documents where they differ
- Hermanos Peral S.L. · F2025-0581 · Scan · Claude Sonnet 5 0/13, GPT-6 Luna 13/13
- Talleres Ibérica, S.A. · PF25/1188 · Clean image · Claude Sonnet 5 0/13, GPT-6 Luna 13/13
- Losa Comunicación S.L. · A-2026-0862 · PDF text · Claude Sonnet 5 0/13, GPT-6 Luna 13/13
- Centro de Formación del Sur S.L. · F2026-0167 · PDF text · Claude Sonnet 5 0/13, GPT-6 Luna 13/13
- Climatización Levante S.L. · F2025-1305 · Scan · Claude Sonnet 5 12/13, GPT-6 Luna 13/13
- Centro de Formación del Sur S.L.U. · F2025-1424 · Scan · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Academia Checa S.L. · F2025-0208 · Clean image · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Gráficas Levante S.L.U. · F2025-0221 · Clean image · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
Invoice extraction · French (97 documents)
GPT-6 Luna is more accurate by 0.2 pp. GPT-6 Luna is 19.5× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 98.7% | 98.9% |
| Error-free documents | 96% | 86% |
| Per 1,000 documents | $11.02 | $0.57 |
| To fix / 1,000 | 41 | 143 |
| Speed (median) | 4.4 sec | 3.8 sec |
| Valid answers | 99.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Supplier | 99.0% | 99.0% |
| Supplier ID (EIN, IČO…) | 99.0% | 99.0% |
| VAT ID | 99.0% | 98.0% |
| Invoice number | 99.0% | 99.0% |
| Payment reference | 97.9% | 100.0% |
| Invoice date | 99.0% | 100.0% |
| Tax point | 99.0% | 100.0% |
| Due date | 99.0% | 100.0% |
| Net by tax rate | 96.9% | 100.0% |
| Tax by rate | 99.0% | 100.0% |
| Total | 99.0% | 100.0% |
| Currency | 99.0% | 100.0% |
| Bank account | 97.9% | 90.8% |
Documents where they differ
- Établissements Bertrand SASU · F2025-0066 · Scan · Claude Sonnet 5 11/13, GPT-6 Luna 13/13
- Nettoyage Mercier SASU · AC-10176 · Clean image · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Carrosserie Gauthier SASU · F-11079 · Scan · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Établissements Denis SASU · FAC-2025-00031 · Scan · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Carrosserie Alain Meunier EI · FAC-2026-00869 · Clean image · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Catherine Dubois EI · FAC-2025-00139 · PDF text · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- SAS Thomas & Fils · FAC-2025-01058 · Clean image · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
- Gilbert Marie – Multiservices (EI) · FAC-2025-00924 · Clean image · Claude Sonnet 5 13/13, GPT-6 Luna 12/13
Email triage · English (US) (150 documents)
Claude Sonnet 5 is more accurate by 0.7 pp. GPT-6 Luna is 29.8× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 98.2% | 97.5% |
| Error-free documents | 84% | 77% |
| Per 1,000 documents | $4.01 | $0.13 |
| To fix / 1,000 | 160 | 227 |
| Speed (median) | 3.8 sec | 2.5 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Category | 100.0% | 100.0% |
| Priority | 97.3% | 98.0% |
| Sentiment | 86.7% | 79.3% |
| Customer name | 100.0% | 100.0% |
| Order number | 100.0% | 100.0% |
| Invoice number | 100.0% | 100.0% |
| Amount | 100.0% | 100.0% |
| Currency | 100.0% | 100.0% |
| Deadline | 100.0% | 100.0% |
Documents where they differ
- Omar Cabrera · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
- Laurie Knight · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
- Cody Baldwin · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
- Scott Davis · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
- Sandra Hogan · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
Email triage · Czech (150 documents)
Claude Sonnet 5 is more accurate by 0.2 pp. GPT-6 Luna is 27.6× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 98.2% | 97.9% |
| Error-free documents | 83% | 81% |
| Per 1,000 documents | $4.91 | $0.18 |
| To fix / 1,000 | 167 | 187 |
| Speed (median) | 4.7 sec | 2.6 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Category | 100.0% | 100.0% |
| Priority | 96.0% | 98.7% |
| Sentiment | 87.3% | 82.7% |
| Customer name | 100.0% | 100.0% |
| Order number | 100.0% | 100.0% |
| Invoice number | 100.0% | 100.0% |
| Amount | 100.0% | 100.0% |
| Currency | 100.0% | 100.0% |
| Deadline | 100.0% | 100.0% |
Documents where they differ
- Lucie Kašparová · Email · Claude Sonnet 5 8/9, GPT-6 Luna 9/9
- Simona Kolářová · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
- Petra Šťastná · Email · Claude Sonnet 5 8/9, GPT-6 Luna 9/9
- Šimon Beran · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
- Vanda Vacková · Email · Claude Sonnet 5 8/9, GPT-6 Luna 9/9
Email triage · German (150 documents)
Claude Sonnet 5 is more accurate by 0.9 pp. GPT-6 Luna is 33.5× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 97.9% | 97.0% |
| Error-free documents | 81% | 73% |
| Per 1,000 documents | $5.11 | $0.15 |
| To fix / 1,000 | 187 | 273 |
| Speed (median) | 4.3 sec | 2.8 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Category | 100.0% | 100.0% |
| Priority | 96.0% | 99.3% |
| Sentiment | 84.7% | 73.3% |
| Customer name | 100.0% | 100.0% |
| Order number | 100.0% | 100.0% |
| Invoice number | 100.0% | 100.0% |
| Amount | 100.0% | 100.0% |
| Currency | 100.0% | 100.0% |
| Deadline | 100.0% | 100.0% |
Documents where they differ
- Rolf-Peter Metz · Email · Claude Sonnet 5 7/9, GPT-6 Luna 8/9
- Friedbert Hoffmann · Email · Claude Sonnet 5 8/9, GPT-6 Luna 9/9
- Sylwia Gerlach · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
- Irmela Warmer · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
- Linda Bähr · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
- Augusta Lange · Email · Claude Sonnet 5 8/9, GPT-6 Luna 9/9
- Margret Löchel · Email · Claude Sonnet 5 9/9, GPT-6 Luna 8/9
Personal data detection · English (US) (150 documents)
Claude Sonnet 5 is more accurate by 0.1 pp. GPT-6 Luna is 33.1× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.7% | 99.6% |
| Error-free documents | 98% | 97% |
| Per 1,000 documents | $3.50 | $0.11 |
| To fix / 1,000 | 20 | 27 |
| Speed (median) | 2.5 sec | 1.8 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Names | 100.0% | 98.7% |
| Email addresses | 100.0% | 100.0% |
| Phone numbers | 100.0% | 100.0% |
| Addresses | 98.0% | 98.7% |
| Dates of birth | 100.0% | 100.0% |
| National IDs | 100.0% | 100.0% |
| Bank accounts | 100.0% | 100.0% |
Documents where they differ
- Justin Martinez · Text · Claude Sonnet 5 7/7, GPT-6 Luna 6/7
- Margaret Mills · Text · Claude Sonnet 5 7/7, GPT-6 Luna 6/7
Personal data detection · Czech (150 documents)
GPT-6 Luna is more accurate by 0.8 pp. GPT-6 Luna is 33.9× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 98.9% | 99.6% |
| Error-free documents | 92% | 97% |
| Per 1,000 documents | $4.06 | $0.12 |
| To fix / 1,000 | 80 | 27 |
| Speed (median) | 3.2 sec | 1.8 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Names | 96.7% | 100.0% |
| Email addresses | 100.0% | 100.0% |
| Phone numbers | 100.0% | 100.0% |
| Addresses | 97.3% | 97.3% |
| Dates of birth | 98.0% | 100.0% |
| National IDs | 100.0% | 100.0% |
| Bank accounts | 100.0% | 100.0% |
Documents where they differ
- Růžena Tichá · Text · Claude Sonnet 5 6/7, GPT-6 Luna 7/7
- Rudolf Konečný · Text · Claude Sonnet 5 6/7, GPT-6 Luna 7/7
- Zikmund Sedláček · Text · Claude Sonnet 5 6/7, GPT-6 Luna 7/7
Personal data detection · German (150 documents)
Claude Sonnet 5 is more accurate by 0.1 pp. GPT-6 Luna is 37.6× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.7% | 99.6% |
| Error-free documents | 98% | 97% |
| Per 1,000 documents | $4.40 | $0.12 |
| To fix / 1,000 | 20 | 27 |
| Speed (median) | 3.7 sec | 2.2 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Names | 98.7% | 100.0% |
| Email addresses | 100.0% | 100.0% |
| Phone numbers | 100.0% | 100.0% |
| Addresses | 100.0% | 97.3% |
| Dates of birth | 99.3% | 100.0% |
| National IDs | 100.0% | 100.0% |
| Bank accounts | 100.0% | 100.0% |
Documents where they differ
- Gunnar Zorbach · Text · Claude Sonnet 5 7/7, GPT-6 Luna 6/7
Contract clauses · English (US) (150 documents)
Claude Sonnet 5 is more accurate by 0.2 pp. GPT-6 Luna is 25.1× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.0% | 99.8% |
| Error-free documents | 100% | 99% |
| Per 1,000 documents | $5.04 | $0.20 |
| To fix / 1,000 | 0 | 13 |
| Speed (median) | 3.4 sec | 2.7 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Provider | 100.0% | 98.7% |
| Provider ID | 100.0% | 100.0% |
| Client | 100.0% | 98.7% |
| Client ID | 100.0% | 100.0% |
| Contract type | 100.0% | 100.0% |
| Price (net) | 100.0% | 100.0% |
| Currency | 100.0% | 100.0% |
| Billing period | 100.0% | 100.0% |
| Effective date | 100.0% | 100.0% |
| End date | 100.0% | 100.0% |
| Notice period (months) | 100.0% | 100.0% |
| Late-payment penalty (% per day) | 100.0% | 100.0% |
| Governing law | 100.0% | 100.0% |
Documents where they differ
- Pittsburgh Construction Co. · Evans Electric LLC · Contract text · Claude Sonnet 5 13/13, GPT-6 Luna 11/13
Contract clauses · Czech (150 documents)
Both are equally accurate. GPT-6 Luna is 22.4× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.0% | 100.0% |
| Error-free documents | 100% | 100% |
| Per 1,000 documents | $6.05 | $0.27 |
| To fix / 1,000 | 0 | 0 |
| Speed (median) | 3.6 sec | 2.6 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Provider | 100.0% | 100.0% |
| Provider ID | 100.0% | 100.0% |
| Client | 100.0% | 100.0% |
| Client ID | 100.0% | 100.0% |
| Contract type | 100.0% | 100.0% |
| Price (net) | 100.0% | 100.0% |
| Currency | 100.0% | 100.0% |
| Billing period | 100.0% | 100.0% |
| Effective date | 100.0% | 100.0% |
| End date | 100.0% | 100.0% |
| Notice period (months) | 100.0% | 100.0% |
| Late-payment penalty (% per day) | 100.0% | 100.0% |
| Governing law | 100.0% | 100.0% |
Documents where they differ
Both models got the same number of fields right on every example document.
Contract clauses · German (150 documents)
GPT-6 Luna is more accurate by 1.8 pp. GPT-6 Luna is 26.7× cheaper.
| Metric | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Correct fields | 98.2% | 100.0% |
| Error-free documents | 87% | 100% |
| Per 1,000 documents | $7.11 | $0.27 |
| To fix / 1,000 | 133 | 0 |
| Speed (median) | 4.3 sec | 3.0 sec |
| Valid answers | 100.0% | 100.0% |
Accuracy per field
| Field | Claude Sonnet 5 | GPT-6 Luna |
|---|---|---|
| Provider | 100.0% | 100.0% |
| Provider ID | 87.3% | 100.0% |
| Client | 100.0% | 100.0% |
| Client ID | 89.3% | 100.0% |
| Contract type | 100.0% | 100.0% |
| Price (net) | 100.0% | 100.0% |
| Currency | 100.0% | 100.0% |
| Billing period | 100.0% | 100.0% |
| Effective date | 100.0% | 100.0% |
| End date | 100.0% | 100.0% |
| Notice period (months) | 100.0% | 100.0% |
| Late-payment penalty (% per day) | 100.0% | 100.0% |
| Governing law | 100.0% | 100.0% |
Documents where they differ
- Wesermetall e.K. · Weserform GmbH · Contract text · Claude Sonnet 5 11/13, GPT-6 Luna 13/13
- Schulz Systemtechnik GmbH · Elbdata GmbH · Contract text · Claude Sonnet 5 11/13, GPT-6 Luna 13/13
- Richter & Albrecht e.K. · Elbmetall UG (haftungsbeschränkt) · Contract text · Claude Sonnet 5 11/13, GPT-6 Luna 13/13
- Meyer AG · Hartmann GmbH · Contract text · Claude Sonnet 5 11/13, GPT-6 Luna 13/13
- Schuster Medien e.K. · Bäckerei Böhm GmbH · Contract text · Claude Sonnet 5 12/13, GPT-6 Luna 13/13