Fireworks AI · featurePrompt caching

Valid from

2025-06-12

Valid until

still holds

Published

2025-06-12

Recorded by AsOf

2026-11-05

Checked

2026-11-05 · its source page states the value and the date

Confidence

0.70

Sources

“On top of fast inference runtime, we’ve enabled prompt caching for vision models , both text and image portions of your prompts can benefit from caching to reduce time to first token by up to 80%.”— what the page says, read 2026-11-05

• Vision Language Models: Llama 4, InternVL3, Caching, LoRAprimary · fireworks.ai · 2025-06-12

/claim/clm_E2E7876831DE.md/v1/claims/clm_E2E7876831DEtimelinesubmit a correction