Krea-2-Turbo

Pricing — 1 offering(s)

Showing the active price and any recorded history. Full pricing history is available via the paid API — see API docs.

Capability profile

Operator guidance

Route here for fast, cheap iteration and open-weight self-hosting. Step up to a frontier model when peak photorealism, text rendering, or compositional accuracy matter more than speed — this entry doesn't have independently verified strength on those dimensions, only a speed/cost claim.

Use cases

• Rapid ideation and concepting where iteration speed matters more than peak fidelity
• High-volume, cost-sensitive generation ($0.008/megapixel)
• Open-weight self-hosting under the Krea 2 Community License

Limitations

• Most capability dimensions beyond inference-speed are unrated — Krea has not published independently verifiable claims for photorealism, text-rendering, prompt-adherence, compositional-accuracy, or a maximum resolution
• Krea's own model card acknowledges prompt-following can be inconsistent depending on prompt style/specificity/language
• No independently published GenEval / T2I-CompBench / benchmark score found. Checked 2026-08-19 (SCO-462 sweep) against the Artificial Analysis Text to Image Arena — no entry named exactly 'Krea 2 Turbo' exists there; the closest candidates ('Krea 2 Medium Turbo', Elo 1216, and 'Krea 2 Large', Elo 1219) can't be confidently matched to this specific checkpoint since neither Krea's own model card nor fal.ai's listing states a Large/Medium size tier for it. Genuinely still unavailable, not just unmatched — don't re-cite either AA figure for this model without a clearer size-tier confirmation from Krea

Citations

• Krea 2 Turbo model card (Krea, Hugging Face)
• Krea 2 Turbo on fal.ai