would be interesting how they compare to gpt-oss-120b. The latter one runs also very fast and pricing is currently much better than qwen3-next on many providers. Would expect that if this model is such fast pricing should be similar or even lower.