Fireworks AI API pricing history

Every Fireworks AI list price we can evidence, back to January 24, 2024 — 14 dated observations across 9 models. Each row links to the page the figure was read off, usually a timestamped snapshot in the Internet Archive. Plus 3 more points recorded by LiteLLM, an aggregator, across 3 of the models above — not seller-verified, marked as such in the tables below.

Price per 1M output tokens over time, logarithmic scale Mixtral 8x7B (MoE size-tier 0B-56B): $0.50 on September 5, 2024 to $0.50 on March 2, 2026; Llama 3.1 405B: $3.00 on September 5, 2024 to $3.00 on September 3, 2025; Llama 70B-class (size-tier more than 16B): $0.90 on March 2, 2026, a single observation; Llama 4 Scout (Basic): $0.60 on September 3, 2025, a single observation; Llama 4 Maverick (Basic): $0.88 on September 3, 2025, a single observation; Llama 3/3.1 8B-class (size-tier 0B-16B): $0.20 on September 5, 2024, a single observation; Llama 3/3.1 70B-class (size-tier 16.1B-80B): $0.90 on September 5, 2024, a single observation; Mixtral 8x7B: $1.60 on January 24, 2024, a single observation $0.10$0.30$1.00$3.00$10.0020252026 Mixtral 8x7B (MoE size-tier 0B-56B) — September 5, 2024 — $0.50 per 1M output tokensMixtral 8x7B (MoE size-tier 0B-56B) — March 4, 2025 — $0.50 per 1M output tokensMixtral 8x7B (MoE size-tier 0B-56B) — September 3, 2025 — $0.50 per 1M output tokensMixtral 8x7B (MoE size-tier 0B-56B) — March 2, 2026 — $0.50 per 1M output tokensLlama 3.1 405B — September 5, 2024 — $3.00 per 1M output tokensLlama 3.1 405B — March 4, 2025 — $3.00 per 1M output tokensLlama 3.1 405B — September 3, 2025 — $3.00 per 1M output tokensLlama 70B-class (size-tier more than 16B) — March 2, 2026 — $0.90 per 1M output tokensLlama 4 Scout (Basic) — September 3, 2025 — $0.60 per 1M output tokensLlama 4 Maverick (Basic) — September 3, 2025 — $0.88 per 1M output tokensLlama 3/3.1 8B-class (size-tier 0B-16B) — September 5, 2024 — $0.20 per 1M output tokensLlama 3/3.1 70B-class (size-tier 16.1B-80B) — September 5, 2024 — $0.90 per 1M output tokensMixtral 8x7B — January 24, 2024 — $1.60 per 1M output tokens
  • Mixtral 8x7B (MoE size-tier 0B-56B)
  • Llama 3.1 405B
  • Llama 70B-class (size-tier more than 16B) (one observation)
  • Llama 4 Scout (Basic) (one observation)
  • Llama 4 Maverick (Basic) (one observation)
  • Llama 3/3.1 8B-class (size-tier 0B-16B) (one observation)
  • Llama 3/3.1 70B-class (size-tier 16.1B-80B) (one observation)
  • Mixtral 8x7B (one observation)
Showing the 8 most-observed models; 1 more are in the dated tables below.

How to read this, and what it is not

  • These are list prices, standard synchronous tier. Batch, cached-input and off-peak discounts are excluded, so real bills are often lower than the line.
  • A date is an observation, not a change date. A snapshot dated March 2024 means the page said that in March 2024 — the price may have taken effect earlier. Rows marked Launch or Price change do carry the vendor's own date.
  • Rows marked "LiteLLM, monthly" are not seller-verified. They are read from a dated commit of LiteLLM's aggregated price file, not from the provider's own page — a weaker guarantee than every other row here. Their date is the month a price was first recorded in that monthly snapshot, never the day it changed, and the true change could fall anywhere in the weeks before it. Each one still links to the exact commit it came from, so you can check it yourself.
  • Gaps are gaps. Lines connect observations; they are not evidence about the days in between, and a missing model means nobody has evidenced it yet, not that its price held.
  • Same name, different thing. Models are quietly revised under a stable name. A flat line can hide a model that got better at the same price — which is itself a price cut per unit of capability, and one this dataset cannot see.
  • Prices in CNY are shown in CNY. Converting an old observation with today's exchange rate would invent a number, so we don't.

Llama 2 70B-class (size-tier 16.1B-80B)

DateInput / 1MOutput / 1MEvidenceLink
$0.70$2.80 Pricing page source

Llama 3.1 405B

DateInput / 1MOutput / 1MEvidenceLink
$3.00$3.00 Launch (blended) source
$3.00$3.00 Pricing page (blended) source
$3.00$3.00 Pricing page (blended) source

Output price unchanged across 363 days of observations.

Llama 3/3.1 70B-class (size-tier 16.1B-80B)

DateInput / 1MOutput / 1MEvidenceLink
$0.90$0.90 Pricing page (blended) source

Llama 3/3.1 8B-class (size-tier 0B-16B)

DateInput / 1MOutput / 1MEvidenceLink
$0.20$0.20 Pricing page (blended) source

Llama 4 Maverick (Basic)

DateInput / 1MOutput / 1MEvidenceLink
$0.22$0.88 First recorded — LiteLLM, monthly source
$0.22$0.88 Launch source

1 row marked LiteLLM, monthly above is recorded by an aggregator, not read from Fireworks AI's own page — see "How to read this" above.

Llama 4 Scout (Basic)

DateInput / 1MOutput / 1MEvidenceLink
$0.15$0.60 First recorded — LiteLLM, monthly source
$0.15$0.60 Launch source

1 row marked LiteLLM, monthly above is recorded by an aggregator, not read from Fireworks AI's own page — see "How to read this" above.

Llama 70B-class (size-tier more than 16B)

DateInput / 1MOutput / 1MEvidenceLink
$0.90$0.90 Pricing page (blended) source

Mixtral 8x7B

DateInput / 1MOutput / 1MEvidenceLink
$0.40$1.60 Pricing page source

Mixtral 8x7B (MoE size-tier 0B-56B)

DateInput / 1MOutput / 1MEvidenceLink
$0.50$0.50 Price change (blended) source
$0.50$0.50 Pricing page (blended) source
$0.50$0.50 Pricing page (blended) source
$0.50$0.50 First recorded — LiteLLM, monthly source
$0.50$0.50 Pricing page (blended) source

Output price unchanged across 543 days of observations.

1 row marked LiteLLM, monthly above is recorded by an aggregator, not read from Fireworks AI's own page — see "How to read this" above.

Back to every provider's price history · take the data: price-history.json (CC BY 4.0).