GLM-5.3 Flash vs GPT-6 Luna
Same documents, same prompt: where each model gets it right, what it really costs, and how fast it responds.
Verdict
GLM-5.3 Flash is more accurate; GPT-6 Luna is cheaper in 30 of 30 combinations.
More accurate
GLM-5.3 Flash
wins 19 of 30
Cheaper
GPT-6 Luna
cheaper in 30 of 30
Faster
GPT-6 Luna
3.0 sec vs 4.4 sec
Fewer fixes
GLM-5.3 Flash
13 vs 33 documents per 1,000
Across 30 task and document-language combinations (speed and fixes: median).
At a glance
| Accuracy (correct fields) | Per 1,000 documents | |||||
|---|---|---|---|---|---|---|
| Task · documents | GLM-5.3 Flash | GPT-6 Luna | Difference | GLM-5.3 Flash | GPT-6 Luna | Difference |
| Invoice extraction | 99.94% | 99.70% | ▼ 0.24 | $0.95 | $0.51 | −46 % |
| Email triage | 99.41% | 97.48% | ▼ 1.93 | $0.33 | $0.13 | −59 % |
| Personal data detection | 100.00% | 99.62% | ▼ 0.38 | $0.33 | $0.11 | −68 % |
| Contract clauses | 99.49% | 99.79% | ▲ 0.3 | $0.40 | $0.20 | −50 % |
| Document summaries | 100.00% | 100.00% | ±0 | $0.82 | $0.16 | −81 % |
| Questions about a document | 100.00% | 99.11% | ▼ 0.89 | $0.32 | $0.15 | −53 % |
| Accuracy (correct fields) | Per 1,000 documents | |||||
|---|---|---|---|---|---|---|
| Task · documents | GLM-5.3 Flash | GPT-6 Luna | Difference | GLM-5.3 Flash | GPT-6 Luna | Difference |
| Invoice extraction | 99.69% | 99.28% | ▼ 0.41 | $1.02 | $0.51 | −51 % |
| Email triage | 98.89% | 97.93% | ▼ 0.96 | $0.52 | $0.18 | −66 % |
| Personal data detection | 99.33% | 99.62% | ▲ 0.29 | $0.49 | $0.12 | −76 % |
| Contract clauses | 100.00% | 100.00% | ±0 | $0.69 | $0.27 | −61 % |
| Document summaries | 100.00% | 100.00% | ±0 | $1.13 | $0.21 | −81 % |
| Questions about a document | 100.00% | 100.00% | ±0 | $0.46 | $0.23 | −50 % |
| Accuracy (correct fields) | Per 1,000 documents | |||||
|---|---|---|---|---|---|---|
| Task · documents | GLM-5.3 Flash | GPT-6 Luna | Difference | GLM-5.3 Flash | GPT-6 Luna | Difference |
| Invoice extraction | 99.54% | 99.23% | ▼ 0.31 | $1.01 | $0.54 | −46 % |
| Email triage | 98.00% | 96.96% | ▼ 1.04 | $0.43 | $0.15 | −64 % |
| Personal data detection | 99.81% | 99.62% | ▼ 0.19 | $0.40 | $0.12 | −70 % |
| Contract clauses | 100.00% | 100.00% | ±0 | $0.65 | $0.27 | −59 % |
| Document summaries | 100.00% | 100.00% | ±0 | $1.08 | $0.19 | −82 % |
| Questions about a document | 98.78% | 98.00% | ▼ 0.78 | $0.41 | $0.20 | −51 % |
| Accuracy (correct fields) | Per 1,000 documents | |||||
|---|---|---|---|---|---|---|
| Task · documents | GLM-5.3 Flash | GPT-6 Luna | Difference | GLM-5.3 Flash | GPT-6 Luna | Difference |
| Invoice extraction | 99.38% | 99.13% | ▼ 0.25 | $1.06 | $0.55 | −48 % |
| Email triage | 99.11% | 96.81% | ▼ 2.3 | $0.58 | $0.16 | −73 % |
| Personal data detection | 99.81% | 99.71% | ▼ 0.1 | $0.46 | $0.12 | −74 % |
| Contract clauses | 100.00% | 100.00% | ±0 | $0.64 | $0.26 | −59 % |
| Document summaries | 96.67% | 97.00% | ▲ 0.33 | $1.54 | $0.19 | −88 % |
| Questions about a document | 100.00% | 99.67% | ▼ 0.33 | $0.36 | $0.19 | −47 % |
| Accuracy (correct fields) | Per 1,000 documents | |||||
|---|---|---|---|---|---|---|
| Task · documents | GLM-5.3 Flash | GPT-6 Luna | Difference | GLM-5.3 Flash | GPT-6 Luna | Difference |
| Invoice extraction | 99.08% | 99.08% | ±0 | $0.98 | $0.56 | −43 % |
| Email triage | 98.52% | 97.26% | ▼ 1.26 | $0.48 | $0.16 | −67 % |
| Personal data detection | 99.81% | 99.62% | ▼ 0.19 | $0.41 | $0.12 | −70 % |
| Contract clauses | 99.74% | 99.33% | ▼ 0.41 | $0.69 | $0.27 | −60 % |
| Document summaries | 100.00% | 98.78% | ▼ 1.22 | $1.25 | $0.20 | −84 % |
| Questions about a document | 99.67% | 99.56% | ▼ 0.11 | $0.39 | $0.20 | −48 % |
Difference: GPT-6 Luna vs. GLM-5.3 Flash; accuracy in percentage points, price in percent. The better value is in bold. Click a task to see the details.
Invoice extraction · English (US) (150 documents)
GLM-5.3 Flash is more accurate by 0.2 pp. GPT-6 Luna is 1.9× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.94% | 99.70% |
| Error-free documents | 99.33% | 96.67% |
| Per 1,000 documents | $0.95 | $0.51 |
| To fix / 1,000 | 7 | 33 |
| Speed (median) | 4.2 sec | 3.4 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Invoice extraction · Czech (150 documents)
GLM-5.3 Flash is more accurate by 0.4 pp. GPT-6 Luna is 2× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.69% | 99.28% |
| Error-free documents | 96.67% | 93.33% |
| Per 1,000 documents | $1.02 | $0.51 |
| To fix / 1,000 | 33 | 67 |
| Speed (median) | 4.5 sec | 3.2 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Invoice extraction · German (150 documents)
GLM-5.3 Flash is more accurate by 0.3 pp. GPT-6 Luna is 1.9× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.54% | 99.23% |
| Error-free documents | 94.67% | 91.33% |
| Per 1,000 documents | $1.01 | $0.54 |
| To fix / 1,000 | 53 | 87 |
| Speed (median) | 4.4 sec | 3.7 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Invoice extraction · Spanish (150 documents)
GLM-5.3 Flash is more accurate by 0.3 pp. GPT-6 Luna is 1.9× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.38% | 99.13% |
| Error-free documents | 92.67% | 88.67% |
| Per 1,000 documents | $1.06 | $0.55 |
| To fix / 1,000 | 73 | 113 |
| Speed (median) | 5.0 sec | 3.9 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Invoice extraction · French (150 documents)
Both are equally accurate. GPT-6 Luna is 1.8× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.08% | 99.08% |
| Error-free documents | 89.33% | 88.00% |
| Per 1,000 documents | $0.98 | $0.56 |
| To fix / 1,000 | 107 | 120 |
| Speed (median) | 4.7 sec | 3.9 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Email triage · English (US) (150 documents)
GLM-5.3 Flash is more accurate by 1.9 pp. GPT-6 Luna is 2.5× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.41% | 97.48% |
| Error-free documents | 94.67% | 77.33% |
| Per 1,000 documents | $0.33 | $0.13 |
| To fix / 1,000 | 53 | 227 |
| Speed (median) | 3.9 sec | 2.5 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Email triage · Czech (150 documents)
GLM-5.3 Flash is more accurate by 1 pp. GPT-6 Luna is 2.9× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 98.89% | 97.93% |
| Error-free documents | 90.00% | 81.33% |
| Per 1,000 documents | $0.52 | $0.18 |
| To fix / 1,000 | 100 | 187 |
| Speed (median) | 4.7 sec | 2.6 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Email triage · German (150 documents)
GLM-5.3 Flash is more accurate by 1 pp. GPT-6 Luna is 2.8× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 98.00% | 96.96% |
| Error-free documents | 82.00% | 72.67% |
| Per 1,000 documents | $0.43 | $0.15 |
| To fix / 1,000 | 180 | 273 |
| Speed (median) | 3.8 sec | 2.8 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Email triage · Spanish (150 documents)
GLM-5.3 Flash is more accurate by 2.3 pp. GPT-6 Luna is 3.7× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.11% | 96.81% |
| Error-free documents | 92.00% | 72.00% |
| Per 1,000 documents | $0.58 | $0.16 |
| To fix / 1,000 | 80 | 280 |
| Speed (median) | 4.5 sec | 3.1 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Email triage · French (150 documents)
GLM-5.3 Flash is more accurate by 1.3 pp. GPT-6 Luna is 3× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 98.52% | 97.26% |
| Error-free documents | 86.67% | 76.00% |
| Per 1,000 documents | $0.48 | $0.16 |
| To fix / 1,000 | 133 | 240 |
| Speed (median) | 4.2 sec | 2.8 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Personal data detection · English (US) (150 documents)
GLM-5.3 Flash is more accurate by 0.4 pp. GPT-6 Luna is 3.1× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 99.62% |
| Error-free documents | 100.00% | 97.33% |
| Per 1,000 documents | $0.33 | $0.11 |
| To fix / 1,000 | 0 | 27 |
| Speed (median) | 2.8 sec | 1.8 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Personal data detection · Czech (150 documents)
GPT-6 Luna is more accurate by 0.3 pp. GPT-6 Luna is 4.1× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.33% | 99.62% |
| Error-free documents | 99.33% | 97.33% |
| Per 1,000 documents | $0.49 | $0.12 |
| To fix / 1,000 | 7 | 27 |
| Speed (median) | 3.8 sec | 1.8 sec |
| Valid answers | 99.33% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Personal data detection · German (150 documents)
GLM-5.3 Flash is more accurate by 0.2 pp. GPT-6 Luna is 3.4× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.81% | 99.62% |
| Error-free documents | 98.67% | 97.33% |
| Per 1,000 documents | $0.40 | $0.12 |
| To fix / 1,000 | 13 | 27 |
| Speed (median) | 3.5 sec | 2.2 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Personal data detection · Spanish (150 documents)
GLM-5.3 Flash is more accurate by 0.1 pp. GPT-6 Luna is 3.9× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.81% | 99.71% |
| Error-free documents | 98.67% | 98.00% |
| Per 1,000 documents | $0.46 | $0.12 |
| To fix / 1,000 | 13 | 20 |
| Speed (median) | 3.4 sec | 2.3 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Personal data detection · French (150 documents)
GLM-5.3 Flash is more accurate by 0.2 pp. GPT-6 Luna is 3.3× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.81% | 99.62% |
| Error-free documents | 98.67% | 97.33% |
| Per 1,000 documents | $0.41 | $0.12 |
| To fix / 1,000 | 13 | 27 |
| Speed (median) | 3.3 sec | 2.4 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Contract clauses · English (US) (150 documents)
GPT-6 Luna is more accurate by 0.3 pp. GPT-6 Luna is 2× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.49% | 99.79% |
| Error-free documents | 96.00% | 98.67% |
| Per 1,000 documents | $0.40 | $0.20 |
| To fix / 1,000 | 40 | 13 |
| Speed (median) | 3.4 sec | 2.7 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Contract clauses · Czech (150 documents)
Both are equally accurate. GPT-6 Luna is 2.6× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 100.00% |
| Error-free documents | 100.00% | 100.00% |
| Per 1,000 documents | $0.69 | $0.27 |
| To fix / 1,000 | 0 | 0 |
| Speed (median) | 4.9 sec | 2.6 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Contract clauses · German (150 documents)
Both are equally accurate. GPT-6 Luna is 2.5× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 100.00% |
| Error-free documents | 100.00% | 100.00% |
| Per 1,000 documents | $0.65 | $0.27 |
| To fix / 1,000 | 0 | 0 |
| Speed (median) | 5.5 sec | 3.0 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Contract clauses · Spanish (150 documents)
Both are equally accurate. GPT-6 Luna is 2.4× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 100.00% |
| Error-free documents | 100.00% | 100.00% |
| Per 1,000 documents | $0.64 | $0.26 |
| To fix / 1,000 | 0 | 0 |
| Speed (median) | 4.9 sec | 3.1 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Contract clauses · French (150 documents)
GLM-5.3 Flash is more accurate by 0.4 pp. GPT-6 Luna is 2.5× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.74% | 99.33% |
| Error-free documents | 96.67% | 91.33% |
| Per 1,000 documents | $0.69 | $0.27 |
| To fix / 1,000 | 33 | 87 |
| Speed (median) | 5.1 sec | 3.2 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Document summaries · English (US) (150 documents)
Both are equally accurate. GPT-6 Luna is 5.2× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 100.00% |
| Error-free documents | 100.00% | 100.00% |
| Per 1,000 documents | $0.82 | $0.16 |
| To fix / 1,000 | 0 | 0 |
| Speed (median) | 7.5 sec | 2.9 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Document summaries · Czech (150 documents)
Both are equally accurate. GPT-6 Luna is 5.3× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 100.00% |
| Error-free documents | 100.00% | 100.00% |
| Per 1,000 documents | $1.13 | $0.21 |
| To fix / 1,000 | 0 | 0 |
| Speed (median) | 10.4 sec | 3.4 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Document summaries · German (150 documents)
Both are equally accurate. GPT-6 Luna is 5.5× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 100.00% |
| Error-free documents | 100.00% | 100.00% |
| Per 1,000 documents | $1.08 | $0.19 |
| To fix / 1,000 | 0 | 0 |
| Speed (median) | 8.4 sec | 3.3 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Document summaries · Spanish (150 documents)
GPT-6 Luna is more accurate by 0.3 pp. GPT-6 Luna is 8.1× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 96.67% | 97.00% |
| Error-free documents | 96.67% | 82.00% |
| Per 1,000 documents | $1.54 | $0.19 |
| To fix / 1,000 | 33 | 180 |
| Speed (median) | 9.2 sec | 3.0 sec |
| Valid answers | 96.67% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Document summaries · French (150 documents)
GLM-5.3 Flash is more accurate by 1.2 pp. GPT-6 Luna is 6.4× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 98.78% |
| Error-free documents | 100.00% | 92.67% |
| Per 1,000 documents | $1.25 | $0.20 |
| To fix / 1,000 | 0 | 73 |
| Speed (median) | 8.7 sec | 3.2 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Questions about a document · English (US) (150 documents)
GLM-5.3 Flash is more accurate by 0.9 pp. GPT-6 Luna is 2.1× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 99.11% |
| Error-free documents | 100.00% | 96.00% |
| Per 1,000 documents | $0.32 | $0.15 |
| To fix / 1,000 | 0 | 40 |
| Speed (median) | 2.4 sec | 3.1 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Questions about a document · Czech (150 documents)
Both are equally accurate. GPT-6 Luna is 2× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 100.00% |
| Error-free documents | 100.00% | 100.00% |
| Per 1,000 documents | $0.46 | $0.23 |
| To fix / 1,000 | 0 | 0 |
| Speed (median) | 3.0 sec | 3.1 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Questions about a document · German (150 documents)
GLM-5.3 Flash is more accurate by 0.8 pp. GPT-6 Luna is 2× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 98.78% | 98.00% |
| Error-free documents | 95.33% | 91.33% |
| Per 1,000 documents | $0.41 | $0.20 |
| To fix / 1,000 | 47 | 87 |
| Speed (median) | 2.3 sec | 3.2 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Questions about a document · Spanish (150 documents)
GLM-5.3 Flash is more accurate by 0.3 pp. GPT-6 Luna is 1.9× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 100.00% | 99.67% |
| Error-free documents | 100.00% | 98.67% |
| Per 1,000 documents | $0.36 | $0.19 |
| To fix / 1,000 | 0 | 13 |
| Speed (median) | 2.3 sec | 2.9 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.
Questions about a document · French (150 documents)
GLM-5.3 Flash is more accurate by 0.1 pp. GPT-6 Luna is 1.9× cheaper.
| Metric | GLM-5.3 Flash | GPT-6 Luna |
|---|---|---|
| Correct fields | 99.67% | 99.56% |
| Error-free documents | 98.67% | 98.00% |
| Per 1,000 documents | $0.39 | $0.20 |
| To fix / 1,000 | 13 | 20 |
| Speed (median) | 2.2 sec | 3.0 sec |
| Valid answers | 100.00% | 100.00% |
Free with login
Accuracy field by field and the documents where the models differ.
Just your email, no password.