Model checks
Claude Opus 5.5, provider by provider. Model →
last audit 18 Nov · epoch 31 · evidence checked on-chain16 of 21
providers answered as the model
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.
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
16
SAME
Failed
5
DIFF
Not tested
3
the audit could not run
By provider
Compared with anthropic/claude-opus-5.5 on OpenRouter · pinned to anthropic · reasoning minimal
| Provider | Differed↑ | × noise | Fail line | Noise | Proof |
|---|---|---|---|---|---|
| STEYGSAME | 1% | 0.5× | 16 under | 1.7% | ↗ |
| GesundAISAME | 1% | 0.5× | 16 under | 1.7% | ↗ |
| Open BirdSAME auditors split: 1 DIFF, 1 SAME | 1% | 1.0× | 14 under | 0.9% | ↗ |
| bartly.eth64.deSAME auditors split: 1 DIFF, 1 SAME | 1% | 1.0× | 14 under | 0.9% | ↗ |
| Lab StudiosSAME | 2% | 1.0× | 15 under | 1.7% | ↗ |
| Inference ReadySAME | 2% | 1.0× | 15 under | 1.7% | ↗ |
| ▲ Apex AntSAME | 2% | 1.0× | 15 under | 1.7% | ↗ |
| LLMartSAME | 3% | 1.5× | 14 under | 1.7% | ↗ |
| Super SeederSAME | 3% | 1.5× | 14 under | 1.7% | ↗ |
| ZLKPro-ApiSAME | 3% | 1.5× | 14 under | 1.7% | ↗ |
| Relay WorksSAME | 3% | 1.5× | 14 under | 1.7% | ↗ |
| PawFlor-SI-Gateway ⚡SAME | 3% | 1.5× | 14 under | 1.7% | ↗ |
| AIXSAME | 3% | 1.7× | 13 under | 1.7% | ↗ |
| AI Inference GatewaySAME | 3% | 2.0× | 13 under | 1.7% | ↗ |
| Unified Rack · DeepSeek 240 tok/sSAME | 3% | 2.0× | 13 under | 1.7% | ↗ |
| klinnSAME | 4% | 2.5× | 12 under | 1.7% | ↗ |
| dealerDIFF | 25% | 15× | +10 over | 1.7% | ↗ |
| Hana Gateway ✅DIFF | 25% | 15× | +13 over | 1.7% | ↗ |
| Vikey SellerDIFF | 39% | 24× | +30 over | 1.7% | ↗ |
| TokenSaleDIFF | 49% | 54× | +35 over | 0.9% | ↗ |
| BestDIFF | 49% | 54× | +39 over | 0.9% | ↗ |
| attacompute.ai not tested · timed out | — | — | — | — | ↗ |
| antseed-aggregator not tested · provider unreachable | — | — | — | — | ↗ |
| Venice.ai Proxy not tested · provider unreachable | — | — | — | — | ↗ |
12 older checks · 3 failed · not counting now
| Provider | Differed↑ | × noise | Fail line | Noise | Proof |
|---|---|---|---|---|---|
| Fire Ant 🔥🐜SAME 7 Nov · expired | 0% | 0.0× | 12 under | 0.0% | ↗ |
| antseed-crimson-swift-476bSAME 7 Nov · expired | 0% | 0.0× | 12 under | 0.0% | ↗ |
| InspectantSAME 7 Nov · expired | 1% | 2.0× | 11 under | 0.0% | ↗ |
| 💬 BanterSAME 7 Nov · expired | 3% | 6.0× | 9 under | 0.0% | ↗ |
| Inference ReadySAME 21 Oct · first round, before points | 5% | 0.9× | 10 under | 4.8% | ↗ |
| antseed-cosmic-gecko-c46dDIFF 7 Nov · expired | 25% | 50× | +13 over | 0.0% | ↗ |
| Edith AIDIFF 21 Oct · first round, before points | 34% | 6.2× | +21 over | 5.5% | ↗ |
| api.contactboxtools.me 17 Nov · 1 DIFF report, no verdict | 59% | 35× | +49 over | 1.7% | ↗ |
| Edith AIDIFF 4 Nov · expired | 65% | 97× | +49 over | 0.7% | ↗ |
| RouteIT 4 Nov · expired · not tested · provider unreachable | — | — | — | — | ↗ |
| Apiport 21 Oct · not tested · provider unreachable | — | — | — | — | ↗ |
| JOI SEED 21 Oct · not tested · provider unreachable | — | — | — | — | ↗ |
Where providers miss
| Subject | Passed providers missed | Failed providers missed |
|---|---|---|
| Biology (chromosome numbers) | 7% | 59% |
| Programming (release years) | 2% | 22% |
| Chinese internet (platform dates) | 13% | 32% |
| Astronomy (periods and distances) | 1% | 26% |
| Medical (drug half-lives) | 5% | 25% |
| Pop culture (box office and ratings) | 1% | 8% |
| Chinese literature (publication years) | 0% | 19% |
| Internet culture (dates and numbers) | 2% | 18% |
| Chinese history (specific dates) | 0% | 16% |
| Chinese geography (measurements) | 4% | 24% |
Audit rounds
| Round | Passed | Failed | Not tested |
|---|---|---|---|
| 17 Novepoch 31 | 16 | 5+1 waiting | 3 |
| 7 Novepoch 30 | 13 | 5 | 6 |
| 4 Novepoch 29 | 14 | 9 | 4 |
| 21 Octepoch 27 · before points | 7 | 4 | 3 |
Audited by approved auditors · evidence on IPFS, hash-checked on-chain.