Live pricing on OneInfer
Pulled from the OneInfer catalog at request time — rates can change. Capture the timestamp with any benchmark or production decision.
| Model | Provider | Input $/1M tokens | Output $/1M tokens |
|---|---|---|---|
| Open | deepinfra | — | — |
| Open | deepinfra | — | — |
| Open | novita | $5.00 | $25.00 |
| Open | groq | $15.00 | $60.00 |
| Open | akashml | $3.70 | $18.70 |
| Open | deepinfra | $3.70 | $18.70 |
| Open | thinkingmachines | $33.00 | $84.00 |
| Open | thinkingmachines | $18.00 | $45.00 |
| Open | groq | $7.50 | $30.00 |
| Open | akashml | $1.80 | $9.00 |
| Open | groq | $7.50 | $30.00 |
| Open | oneinfer | $125.00 | $425.00 |
| Open | deepinfra | $NaN | $NaN |
| Open | deepinfra | $NaN | $NaN |
| closed models for GLM-5.3 buyers | zai | $140.00 | $440.00 |
Decision snapshot
| Decision factor | GLM-5.3 | closed models |
|---|---|---|
| Text and agent workflows | Evaluate | Evaluate |
| Price | Use current live price | Use current live price |
| Deployment control | Verify weight/license status | Verify current terms |
| Non-text modalities | Verify provider capabilities | Verify provider capabilities |
Benchmark rules
- Compare only the same evaluation and harness version.
- Label vendor-reported results.
- Record token budget and tool policy.
- Do not infer production reliability from one benchmark.
Ready to test the workflow?
Strengths and tradeoffs
Select the model against a representative prompt set, latency target, output budget, tool-calling requirements, and data-control constraints.
Workload recommendation
| Workload | How to choose |
|---|---|
| Budget-sensitive coding | Compare task success per dollar. |
| Long-context analysis | Test retrieval and citation accuracy. |
| Multimodal input | Choose a model/provider that explicitly supports it. |
| Regulated data | Review retention, residency, and deployment terms. |