Model checks
ZLKPro-Api, as the auditors found it. Provider →
last audit 18 Nov · evidence checked on-chain×1.02
on its $ANTS points now
Approved auditors check that each provider serves the model it sells.
- 🧪They buy from it like any buyer: ~100 obscure number questions, compared with the real model.
- 📏The real model misses some too: that is the noise. A provider fails only if it misses clearly more, and two auditors must agree.
- 💰Each pass +2% on $ANTS points, up to +50%. Fails cost more: 1 −4%, 2 −16%, 3 −36%. Checked again every week.
How this works (Antseed's announcement ↗)
Approved auditors check that each provider serves the model it sells.
- 🧪They buy from it like any buyer: ~100 obscure number questions, compared with the real model.
- 📏The real model misses some too: that is the noise. A provider fails only if it misses clearly more, and two auditors must agree.
- 💰Each pass +2% on $ANTS points, up to +50%. Fails cost more: 1 −4%, 2 −16%, 3 −36%. Checked again every week.
Passed
9
SAME · counting now
Last audit
18 Nov
epoch 31 · counts to 32
| Model | Differed↑ | × noise | Fail line | Noise | Proof |
|---|
| Gemini 4 ArgonSAME | 2% | 2.0× | 12 under | 1.0% | ↗ |
| GPT-6 AstraSAME auditors split: 1 DIFF, 1 SAME | 2% | 2.0× | 12 under | 1.0% | ↗ |
| Claude Opus 5.5SAME | 3% | 1.5× | 14 under | 1.7% | ↗ |
| Claude Fable 5.1SAME auditors split: 1 DIFF, 1 SAME | 5% | 1.0× | 20 under | 5.0% | ↗ |
| Claude Sonnet 5.5SAME | 5% | 1.2× | 18 under | 4.6% | ↗ |
| Kimi K3.1 ProSAME | 6% | 2.7× | 11 under | 2.3% | ↗ |
| GLM 5.3SAME | 12% | 1.1× | 25 under | 11.3% | ↗ |
| MiniMax M3SAME | 12% | 0.9× | 31 under | 14.1% | ↗ |
| MiMo V2.6 ProDIFF | 28% | 2.8× | +20 over | 10.0% | ↗ |
| Gemini 3.1 Pro PreviewDIFF | 34% | 8.5× | +26 over | 4.0% | ↗ |
Evidence of some checks not found
Gemini 3.1 Pro PreviewDIFF
32 of 220 differed over 2 sets (fails from 24) · 51 of 150 differed (fails from 25) · the real model differs from itself on 4%
Where it missed
Internet culture (dates and numbers)5/12
Chemistry (boiling points)4/11
Chinese history (specific dates)6/17
Astronomy (periods and distances)10/29
For example
Tarantula Nebula (30 Doradus) estimated total mass (millions of solar masses)real 1 · provider 450
sebelipase alfa elimination half life hoursreal 0.1 · provider 2
The number of chapters (juan) in the original Shishuo Xinyu (A New Account of the Tales of the World)real 3 · provider 36
✓Reasoning switched off with enable_thinking false, as the real model was asked.
MiMo V2.6 ProDIFF
79 of 279 differed over 2 sets (fails from 59) · 32 of 128 differed (fails from 28) · the real model differs from itself on 10%
Where it missed
Biology (chromosome numbers)5/11
Programming (release years)8/19
Cryptography (parameters)6/16
Medical (drug half-lives)3/9
For example
thermal conductivity of silicon carbide at 300K W per mKreal 120 · provider 490
number of times the pineapple is visible in the hawaii five o opening creditsreal 1 · provider 3
1,1,1,3,3-pentafluoropropane bp C at 1atmreal -16 · provider 15
🧠Kept reasoning when asked not to: about 1 reasoning tokens on each of 4 ways to switch it off.
2 older checks · not counting now
| Model | Differed↑ | × noise | Fail line | Noise | Proof |
|---|
| Kimi K3SAME 7 Nov · replaced | 7% | 0.8× | 20 under | 8.7% | ↗ |
| GPT-6.1 Sol 17 Nov · 1 DIFF report, no verdict | 42% | 21× | +25 over | 2.0% | ↗ |
| Round | Passed | Failed | Not tested | Multiplier |
|---|
| 17 Novepoch 31 | 8 | 2+1 waiting | 0 | ×0.82 |
| 7 Novepoch 30 | 8 | 3 | 0 | ×0.80 |
| 4 Novepoch 29 | 0 | 8 | 1 | ×0 |
| 21 Octepoch 27 · before points | 1 | 5 | 0 | — |