Model
Host comparison
Host comparison for Self-hosted (self-hosted). Latest input $0.28 / MTok, output $0.28 / MTok, version 2026-09-01.
Wave 4Open weight
Requests here
12
Sample cost
$0.002749
Baseline
$0.002749
Context
128,000
In your Auto router
Auto would have saved you about $4,204.93 on Llama 3.3 70B last month, by sending Clinical's requests to Gemini 2.5 Flash. An estimate from your matched-quality data, never a ledger line.
| Team | Status | ||
|---|---|---|---|
| Operations | default | — | Being checked (21 checks) |
Prices are versioned. Older versions stay in the book and are never overwritten.
| Version | Host | Input / M | Cached / M | Output / M | Batch |
|---|---|---|---|---|---|
| prices 2026-06-01 | self-hosted | $0.4 | $0.04 | $0.4 | — |
| prices 2026-09-01 | self-hosted | $0.28 | $0.028000000000000004 | $0.28 | 50% |