Explorer ยท 12 models
Find the model that fits the budget
Narrow the runs with the filters, then switch the chart between latency and accuracy. Every view reads the same dataset.
Compare providers
Latency: Delta 645 ms, Harbor Labs 920 ms, Northwind 1.1 s, Cedar AI 1.2 s, Ember 1.2 s
Chart data table
| Provider | Latency |
|---|---|
| Northwind | 1.1 s |
| Harbor Labs | 920 ms |
| Cedar AI | 1.2 s |
| Delta | 645 ms |
| Ember | 1.2 s |
Accuracy: Northwind 90%, Cedar AI 88.3%, Harbor Labs 87%, Delta 86%, Ember 86%
Chart data table
| Provider | Accuracy |
|---|---|
| Northwind | 90% |
| Harbor Labs | 87% |
| Cedar AI | 88.3% |
| Delta | 86% |
| Ember | 86% |
Runs
Evaluation runs
| Model | Provider | Tier | Cost / 1M | Latency | Accuracy | Status |
|---|---|---|---|---|---|---|
| atlas-large | Northwind | large | $2.40 | 1.8 s | 94% | pass |
| atlas-mini | Northwind | small | $0.30 | 420 ms | 86% | pass |
| beacon-7b | Harbor Labs | small | $0.20 | 380 ms | 81% | fail |
| beacon-13b | Harbor Labs | medium | $0.60 | 760 ms | 88% | pass |
| beacon-70b | Harbor Labs | large | $1.90 | 1.6 s | 92% | flaky |
| cedar-lite | Cedar AI | small | $0.15 | 310 ms | 79% | pass |
| cedar-pro | Cedar AI | medium | $0.90 | 980 ms | 90% | pass |
| cedar-max | Cedar AI | large | $3.10 | 2.2 s | 96% | pass |
| delta-coder | Delta | medium | $0.70 | 840 ms | 89% | flaky |
| delta-chat | Delta | small | $0.25 | 450 ms | 83% | pass |
| ember-xl | Ember | large | $2.80 | 2 s | 95% | pass |
| ember-s | Ember | small | $0.18 | 290 ms | 77% | fail |