← All humanizers
GPTinf
GPTinf has a current composite of 35.63/100 in our shared November evaluation. Readability is 71.7 and rescaled meaning preservation is 41.4. 231 of 231 detector measurements are available.
These are the service’s 33 published historical rewrites, re-scored under the same measurement profiles as our open-weight outputs. They do not establish how its current service would perform.
Overall score35.63/100
Panel bypass score37.7%
Readability71.7%
Prompts tested33
Detector human-likelihood scores
Mean reported human score on measured outputs. Missing results stay missing. Coverage is shown beside each detector.
Performance by category
HumanizeBench v2 per-category values.
Quality and penalties
0.8518Raw meaning cosine
0/33Identical outputs
0/33Refusals
10.0Penalty points
| Rule | Occurrences | Deduction |
|---|---|---|
| Meaning drift | 15 | −10 |
Artifact & evaluation record
- Entry
- gptinf
- Source
- Published historical service outputs, newly re-scored
- Family
- Closed implementation · underlying model unknown
- Scoring
- 56.5% bypass · 24% meaning · 12% readability · 7.5% consistency, minus fixed penalties.
- Measurement notes
- NVIDIA quality substitutes; current Polygraph routes; no historical-score fallback.
Inspect the recorded text
The source and rewritten text are available for every sample.
| Sample | Category | Source model | Input words | Readability | Detector scores | |
|---|---|---|---|---|---|---|
| S01 Academic essay | Academic essay | gpt-5.6-terra | 315 | 70.0 | 7/7 | |
| S02 Academic essay | Academic essay | claude-sonnet-5 | 296 | 76.0 | 7/7 | |
| S03 Application essay | Application essay | gemini-3.5-flash | 291 | 70.0 | 7/7 | |
| S04 Application essay | Application essay | gpt-5.6-luna | 271 | 70.0 | 7/7 | |
| S05 Blog post | Blog post | claude-haiku-4-5-20251001 | 280 | 56.7 | 7/7 | |
| S06 Blog post | Blog post | gemini-2.5-flash | 329 | 70.0 | 7/7 | |
| S07 Business email | Business email | gpt-5.6-terra | 262 | 82.0 | 7/7 | |
| S08 Marketing copy | Marketing copy | claude-sonnet-5 | 299 | 73.3 | 7/7 | |
| S09 Marketing copy | Marketing copy | gemini-3.5-flash | 191 | 70.0 | 7/7 | |
| S10 Discussion board | Discussion board | gpt-5.6-luna | 268 | 68.7 | 7/7 | |
| S11 News article | News article | claude-haiku-4-5-20251001 | 300 | 68.7 | 7/7 | |
| S12 Academic essay | Academic essay | gemini-2.5-flash | 259 | 70.0 | 7/7 | |
| S13 Academic essay | Academic essay | gpt-5.6-terra | 349 | 68.3 | 7/7 | |
| S14 Application essay | Application essay | claude-sonnet-5 | 314 | 80.0 | 7/7 | |
| S15 Application essay | Application essay | gemini-3.5-flash | 292 | 68.3 | 7/7 | |
| S16 Blog post | Blog post | gpt-5.6-luna | 313 | 76.0 | 7/7 | |
| S17 Blog post | Blog post | claude-haiku-4-5-20251001 | 281 | 70.0 | 7/7 | |
| S18 Business email | Business email | gemini-2.5-flash | 188 | 80.0 | 7/7 | |
| S19 Marketing copy | Marketing copy | gpt-5.6-terra | 257 | 73.3 | 7/7 | |
| S20 Marketing copy | Marketing copy | claude-sonnet-5 | 191 | 73.3 | 7/7 | |
| S21 Discussion board | Discussion board | gemini-3.5-flash | 251 | 80.0 | 7/7 | |
| S22 News article | News article | gpt-5.6-luna | 298 | 76.0 | 7/7 | |
| S23 Academic essay | Academic essay | claude-haiku-4-5-20251001 | 312 | 50.0 | 7/7 | |
| S24 Academic essay | Academic essay | gemini-2.5-flash | 258 | 70.0 | 7/7 | |
| S25 Application essay | Application essay | gpt-5.6-terra | 300 | 70.0 | 7/7 | |
| S26 Application essay | Application essay | claude-sonnet-5 | 329 | 62.7 | 7/7 | |
| S27 Blog post | Blog post | gemini-3.5-flash | 298 | 75.0 | 7/7 | |
| S28 Blog post | Blog post | gpt-5.6-luna | 323 | 73.3 | 7/7 | |
| S29 Business email | Business email | claude-haiku-4-5-20251001 | 228 | 80.0 | 7/7 | |
| S30 Marketing copy | Marketing copy | gemini-2.5-flash | 277 | 70.0 | 7/7 | |
| S31 Marketing copy | Marketing copy | gpt-5.6-terra | 195 | 75.3 | 7/7 | |
| S32 Discussion board | Discussion board | claude-sonnet-5 | 252 | 70.0 | 7/7 | |
| S33 News article | News article | gemini-3.5-flash | 295 | 78.0 | 7/7 |