GoKawiilUltramarine 1 Flash launches at $0.13 per task, matching Fable-class scores
Ultramarine 1 Flash is a new model priced at $0.13 per task on internal benchmark PI Knowledge Work, which the company says correlates strongly with the Artificial Analysis Intelligence Index. It is built on GLM-5.3 Flash and generates 6.5% fewer tokens than that base model while outscoring Claude Fable 5 (max) and GPT-5.6 Sol (max), which cost $8.19 and $2.09 per task respectively. Pricing on the Ultramarine API is $0.15 per million input tokens, $0.03 for cached input, and $0.50 for output.
GoKawiil's interpretation of the reporting above, not reported fact.
The comparison suggests smaller, cheaper models can approach the performance of far more expensive 'max reasoning' configurations on knowledge-work tasks like finance, law and consulting, which could pressure pricing across the industry. Because the benchmark is internally built, the strong correlation claim with an external index is the company's own framing and would benefit from independent verification.
- Ultramarine 1 Flash costs $0.13 per task, roughly 60x cheaper than Claude Fable 5 (max) at $8.19
- It is built on GLM-5.3 Flash but produces 6.5% fewer tokens per task
- Its scores come close to models costing 50x more, per the company's internal PI Knowledge Work benchmark
Source: paperinstruments.com — Paper Instruments, 2026-10-15
Published there as: “Ultramarine 1 Flash”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.