Provider comparison
gpt-oss-120b
A sourced endpoint-by-endpoint view. Pricing is comparable; performance claims are withheld until independent probes exist.
| Rank by input | Provider | Input / 1M | Output / 1M | Quant. | Endpoint | Evidence |
|---|---|---|---|---|---|---|
| 01 | CoreWeave | $0.030 | $0.17 | fp4 | coreweave/fp4 | Provider-reported |
| 02 | DeepInfra | $0.037 | $0.17 | bf16 | deepinfra/bf16 | Provider-reported |
| 03 | Novita | $0.050 | $0.25 | fp4 | novita/fp4 | Provider-reported |
| 04 | DigitalOcean | $0.055 | $0.385 | unknown | digitalocean | Provider-reported |
| 05 | Google Vertex AI | $0.090 | $0.36 | unknown | google-vertex/global | Provider-reported |
Sorted only by listed input price. No quality, latency, throughput, or reliability equivalence is implied.
Questions
gpt-oss-120b provider comparison questions
Which provider is cheapest for gpt-oss-120b?+
CoreWeave has the lowest provider-reported input price in this snapshot at $0.030 per million tokens. This is not an executable quote.
Which provider is fastest for gpt-oss-120b?+
There is no approved independent latency dataset yet, so this comparison does not rank providers by speed.
Are these prices directly comparable?+
They use the same per-million-token unit and currency, but provider implementation, quantization, tier, and policy may differ. Those dimensions remain visible.
How current is this comparison?+
It is a dated source snapshot and is labeled non-live until the production collectors are activated.
How will independent performance be compared?+
Standardized disclosed-region probes will record TTFT, throughput, end-to-end latency, success/error rates, effective cost, workload, request parameters, and returned model version. Read the methodology →