Across 40 shared benchmarks, GPT-5.6 Sol scores higher on 37 and Inkling on 3. The widest gap is critpt, where GPT-5.6 Sol scores 32.3 against 5.4. Inkling is the cheaper of the two on tracked API pricing ($1.00 against $4.00 per million input tokens).
Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing. Quoted rates are the price-setter row we currently track for each model — its direct or vendor-official listing where one exists (direct, direct), otherwise the lowest tracked offer.