Fair comparison
This table reports a dated internal comparison, not a universal ranking. The combination of quality, semantic fidelity, and formal safety matters more than one detector score.
Internal, versioned benchmark · Updated July 28, 2026. These results describe this test run. They do not guarantee the same outcome for every text, language, or future product version.
Updated July 28, 2026. Ten cases per provider; RewriteAI produced nine usable outputs.
A horizontal detail view on an explicitly truncated 40-to-100 point scale. Exact values and every additional metric remain available in the table below.
85.4
Quality · Certentis
Quality scale 40–100 · truncated axis
| Provider | Cases | Quality | Semantics | Exact citations | Correct language | Word ratio | Human + Mixed |
|---|---|---|---|---|---|---|---|
| Certentis | 10 | 85.4 | 91.7 | 8/10 | 10/10 | 0.98 | 92.96% |
| WriteHuman Academic | 10 | 82.4 | 85.4 | 4/10 | 10/10 | 1.01 | 50.3% |
| StealthGPT Heavy PhD | 10 | 70.0 | 72.8 | 6/10 | 10/10 | 0.82 | 68.75% |
| RewriteAI | 9 | 69.67 | 73.0 | 4/9 | 9/9 | 1.08 | 78.41% |
| Humbot Advanced | 10 | 65.8 | 71.5 | 0/10 | 10/10 | 1.10 | 66.89% |
| Rephrasy v3 | 10 | 52.2 | 49.8 | 0/10 | 10/10 | 1.08 | 82.98% |
| HumanizerPro Advanced | 10 | 42.7 | 62.0 | 2/10 | 6/10 | 1.03 | 93.31% |
Every section uses the same documented cases, showing where a provider stays close to Certentis and where quality, meaning, citations, or language diverge.
WriteHuman was the closest competitor on overall quality. Both kept the correct language in all ten cases, while Certentis led on semantics and exact citations and WriteHuman produced the lower Human-plus-Mixed score.
Quality lead: +3.0
Read the full comparison: WriteHuman Academic →| Metric | Certentis | WriteHuman Academic |
|---|---|---|
| Quality | 85.4 | 82.4 |
| Semantics | 91.7 | 85.4 |
| Exact citations | 8/10 | 4/10 |
| Correct language | 10/10 | 10/10 |
| Word ratio | 0.98 | 1.01 |
| Human + Mixed | 92.96% | 50.3% |
StealthGPT kept the language stable and preserved six of ten citation sets. Its outputs were considerably shorter on average, and both overall quality and semantic fidelity trailed Certentis.
Quality lead: +15.4
Read the full comparison: StealthGPT Heavy PhD →| Metric | Certentis | StealthGPT Heavy PhD |
|---|---|---|
| Quality | 85.4 | 70.0 |
| Semantics | 91.7 | 72.8 |
| Exact citations | 8/10 | 6/10 |
| Correct language | 10/10 | 10/10 |
| Word ratio | 0.98 | 0.82 |
| Human + Mixed | 92.96% | 68.75% |
RewriteAI returned nine rather than ten usable results and stayed close to the source length. Quality, semantics, and citation preservation were nevertheless substantially weaker in the direct comparison.
Quality lead: +15.73
Read the full comparison: RewriteAI →| Metric | Certentis | RewriteAI |
|---|---|---|
| Quality | 85.4 | 69.67 |
| Semantics | 91.7 | 73.0 |
| Exact citations | 8/10 | 4/9 |
| Correct language | 10/10 | 9/9 |
| Word ratio | 0.98 | 1.08 |
| Human + Mixed | 92.96% | 78.41% |
Humbot preserved the language in every case and expanded texts moderately on average. It did not preserve a single complete citation set exactly, while quality and semantics remained well behind Certentis.
Quality lead: +19.6
Read the full comparison: Humbot Advanced →| Metric | Certentis | Humbot Advanced |
|---|---|---|
| Quality | 85.4 | 65.8 |
| Semantics | 91.7 | 71.5 |
| Exact citations | 8/10 | 0/10 |
| Correct language | 10/10 | 10/10 |
| Word ratio | 0.98 | 1.10 |
| Human + Mixed | 92.96% | 66.89% |
Rephrasy achieved a high Human-plus-Mixed score but lost substantial semantic fidelity and preserved no citation set exactly. This is why a detector score alone cannot establish text quality.
Quality lead: +33.2
Read the full comparison: Rephrasy v3 →| Metric | Certentis | Rephrasy v3 |
|---|---|---|
| Quality | 85.4 | 52.2 |
| Semantics | 91.7 | 49.8 |
| Exact citations | 8/10 | 0/10 |
| Correct language | 10/10 | 10/10 |
| Word ratio | 0.98 | 1.08 |
| Human + Mixed | 92.96% | 82.98% |
HumanizerPro recorded the highest Human-plus-Mixed score in the run. At the same time, it had the weakest overall quality and language stability, with only two exact citation sets.
Quality lead: +42.7
Read the full comparison: HumanizerPro Advanced →| Metric | Certentis | HumanizerPro Advanced |
|---|---|---|
| Quality | 85.4 | 42.7 |
| Semantics | 91.7 | 62.0 |
| Exact citations | 8/10 | 2/10 |
| Correct language | 10/10 | 6/10 |
| Word ratio | 0.98 | 1.03 |
| Human + Mixed | 92.96% | 93.31% |
Certentis scored 85.4, ahead of the six compared configurations on this dataset.
One provider reached 93.31% Human + Mixed but scored 42.7 for overall quality and preserved the language in only 6 of 10 cases.
No provider achieved 10 of 10 exact citation sets. Certentis led this run with 8 of 10.
Certentis Evidence
Results, test rules, and limitations are kept separate so it remains clear what was measured and what was not.
