Model checks
MiMo V2.6 Pro, seller by seller. Model →
last audit 7 Nov · epoch 30 · evidence checked on-chain1 of 5
sellers answered as the model
How this works (Antseed's announcement ↗)
Approved auditors check that each seller 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 seller 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 seller 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 seller 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
1
SAME
Failed
4
DIFF
Not tested
1
the audit could not run
By seller
Compared with xiaomi/mimo-v2.6-pro on OpenRouter · pinned to xiaomi/fp8 · reasoning none
| Seller | Differed↑ | × noise | Fail line | Noise | Proof |
|---|---|---|---|---|---|
| Vikey SellerSAME | 10% | 0.8× | 21 under | 11.5% | ↗ |
| LLMartDIFF | 30% | 3.3× | +25 over | 8.8% | ↗ |
| ZLKPro-ApiDIFF | 31% | 3.5× | +29 over | 8.8% | ↗ |
| BestDIFF | 39% | 3.6× | +14 over | 10.8% | ↗ |
| Hana Gateway ✅DIFF | 42% | 3.9× | +17 over | 10.8% | ↗ |
| Relay Works not tested · seller unreachable | — | — | — | — | ↗ |
Where sellers miss
| Subject | Passed sellers missed | Failed sellers missed |
|---|---|---|
| Chinese geography (measurements) | 8% | 18% |
| Chinese internet (platform dates) | 0% | 16% |
| Pop culture (box office and ratings) | 20% | 46% |
| Physics (material properties) | 20% | 29% |
| Chinese history (specific dates) | 0% | 44% |
| Astronomy (periods and distances) | — | 33% |
| Programming (release years) | 0% | 56% |
| Chinese literature (publication years) | — | 29% |
| Internet culture (dates and numbers) | 0% | 32% |
| Chemistry (melting points) | 17% | 52% |
Audited by approved auditors · evidence on IPFS, hash-checked on-chain.