Fireworks AI
Cloud inference platform serving open-weights AI models through a pay-as-you-go API.
HN commenters like Fireworks as a fast, low-cost way to run open-weights models12345, with cheap cached tokens and US-based zero data retention.678 Complaints are scattered, with recent ones about slower new-model availability and deprecated models910, plus a billing and support dispute and occasional reliability issues.1112
Why people recommend it
Speed · 3
Commenters describe inference as really fast, without rate limits in daily use.
“I use it all the time through Fireworks.”
“with the Fireworks.ai API and properly configured OpenCode to be very good”
Show 1 more
Low cost · 3
Users report small bills, cheap per-token pricing, and cache hits billed at 20%.
“On Fireworks AI, for example, if it hits the cache, you pay only 20%.”
“I pay as I go on FireWorks.ai for open model inferencing and no matter how much I use this service my monthly bill is between $10 and $40”
Show 1 more
“I use it on fireworks which is US/ZDR and pretty reliable.”
Open-weights access · 3
Used as a host for open models like Kimi, GLM and DeepSeek, sometimes alongside Together.ai.
“I’ve mostly settled on using a mixture of open weights models through Together.ai and Fireworks.ai”
“I used it unquantized through Fireworks”
Show 1 more
“Opencode+Fireworks+Kimi2.5 and git worktrees mainly”
Production use · 3
Several commenters run production apps or user-facing platforms on it.
“I use Fireworks for a production app.”
“I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform”
Show 1 more
“Fireworks is a great provider but we still ran into issues here and there”
Data retention · 2
Chosen as a US provider offering zero data retention.
“we use Fireworks.ai, and American company with ZDR”
“I use it on fireworks which is US/ZDR and pretty reliable.”
Caveats and dissent
New model lag · 1
One user says new models now arrive slowly and must be requested on Discord.
“Fireworks is not keeping up with new models the last few months.”
Model deprecations · 2
Users note deprecated on-demand Flux models and a single, expensive reranking model.
“I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform”
“I use Fireworks for a production app.”
Billing and support · 1
One commenter was charged during a supposedly free trial and got no email responses.
Occasional issues · 1
A production user hit issues here and there and moved to OpenAI.
“Fireworks is a great provider but we still ran into issues here and there”
Ember-1 quality · 1
One user says the Ember-1 model hallucinates heavily.
“Fireworks is not keeping up with new models the last few months.”
Often compared with
Who recommends it in their own words
Threads where it came up
- MiniMax M3.1 Flash 1 mention
- DeepSeek-v4.1-Flash-1017 1 mention
- Muse Spark Open 1 mention
- Ask HN: Who wants to be hired? (October 2026) 1 mention
- Sonnet 5.5 1 mention
- Grok 4.7 1 mention
Also written as Fireworks, Fireworks.ai. Data: fireworks-ai.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 20
“Fireworks is not keeping up with new models the last few months.”
“I'm off Fireworks.”
“But is really fast (fireworks.ai).”
“To iterate quickly, I rely on an inference provider such as Fireworks.”
“we use Fireworks.ai, and American company with ZDR”
“I use it on fireworks which is US/ZDR and pretty reliable.”
“I use Fireworks for a production app.”
“I pay as I go on FireWorks.ai for open model inferencing and no matter how much I use this service my monthly bill is between $10 and $40”
“I’ve mostly settled on using a mixture of open weights models through Together.ai and Fireworks.ai”
“There have been plenty of other applications (Corel, Paint Shop Pro, Fireworks, Krita etc) with their own workflows and keybindings that weren't confusing for users.”
“hardwired to use deepseek-v4-flash-0731 via FireWorks.ai”
“On Fireworks AI, for example, if it hits the cache, you pay only 20%.”
“I buy tokens mostly from FireWorks.ai in the US”
“Fireworks is a great provider but we still ran into issues here and there”
“I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform”
“I use it all the time through Fireworks.”
“I used it unquantized through Fireworks”
“with the Fireworks.ai API and properly configured OpenCode to be very good”
“I use the US company Fireworks.ai”
“Opencode+Fireworks+Kimi2.5 and git worktrees mainly”