anthropic
synthetic-model · DEMO / USD
- Mean latency
- 620 ms
- Reliability
- 100%
- Exact-match quality
- 100%
- Mock settled cost
- $0.000375
Fictional results · Not a live performance or price claim
Seabiza · Control Center
Model benchmarks
Fixed synthetic suite · Fictional mock evidence · No provider calls
DEMO
Suite v1 · 3 identical tasks · Pass 1
cheaperinference
synthetic-model · DEMO / USD
- Mean latency
- 980 ms
- Reliability
- 67%
- Exact-match quality
- 67%
- Mock settled cost
- Unresolved
Unknown liability retained · Fictional results · Not a live performance or price claim
synthetic-model · DEMO / USD
- Mean latency
- 890 ms
- Reliability
- 100%
- Exact-match quality
- 100%
- Mock settled cost
- $0.000600
Fictional results · Not a live performance or price claim
openai
synthetic-model · DEMO / USD
- Mean latency
- 530 ms
- Reliability
- 100%
- Exact-match quality
- 100%
- Mock settled cost
- $0.000300
Fictional results · Not a live performance or price claim
openrouter
synthetic-model · DEMO / USD
- Mean latency
- 710 ms
- Reliability
- 100%
- Exact-match quality
- 100%
- Mock settled cost
- $0.000450
Fictional results · Not a live performance or price claim
together
synthetic-model · DEMO / USD
- Mean latency
- 800 ms
- Reliability
- 100%
- Exact-match quality
- 100%
- Mock settled cost
- $0.000525
Fictional results · Not a live performance or price claim
Synthetic task set
arithmetic-v1Return only the integer: 17 + 25
classification-v1Return only the category of apple: fruit or vehicle
json-v1Return exactly this JSON: {"safe":true}
75 input / 15 output tokens per permutation are fictional fixtures. Errors count against reliability and quality. Unknown costs are not zero. No currency conversion.