{"slug": "fireworks-ai", "name": "Fireworks AI", "category": "ai", "category_name": "AI models and assistants", "description": "Cloud inference platform serving open-weights AI models through a pay-as-you-go API.", "recommenders": 9, "commenters": 14, "warners": 2, "score": 6.779, "rising": 4.69, "first": "2026-06-05", "latest": "2026-10-19", "why": "HN commenters like Fireworks as a fast, low-cost way to run open-weights models, with cheap cached tokens and US-based zero data retention.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 8, "neutral": 8, "compare": 0, "warn": 2}, "spark": [0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 2, 0], "listed": true, "loved": ["Speed", "Low cost"], "caveat": "New model lag", "short_quote": "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", "icon": true, "domain": "fireworks.ai", "github_repo": null, "official_url": "https://fireworks.ai", "pricing": "paid", "aliases": ["Fireworks AI", "Fireworks", "Fireworks.ai"], "verdict": [{"text": "HN commenters like Fireworks as a fast, low-cost way to run open-weights models", "cites": [{"id": 781566001, "item_id": 48847875, "url": "https://news.ycombinator.com/item?id=48847875", "by": "pimeys", "date": "2026-07-09", "stance": "recommend", "quote": "I use it all the time through Fireworks.", "excerpt": "I use it all the time through Fireworks. The normal version when I pay it myself and the fast one when company pays. It's really fast and I never get rate limited with my daily use.", "context": "", "use_case": "inference hosting", "switched_from": "", "story": "Introducing Muse Spark 1.1"}, {"id": 779098850, "item_id": 48693678, "url": "https://news.ycombinator.com/item?id=48693678", "by": "mark_l_watson", "date": "2026-06-26", "stance": "recommend", "quote": "with the Fireworks.ai API and properly configured OpenCode to be very good", "excerpt": "Your experience with DeepSeek v4 Flash differs from mine: while I usually use DeepSeek v4 Pro (that is also inexpensive), I find using DeepSeek v4 Flash with the Fireworks.ai API and properly configured OpenCode to be very good for routine work, and it is pleasantly very fast. Admittedly I use DeepSeek v4 Pro for difficult problems. I encourage people to at least once a month to do a quick evaluation with their own p", "context": "", "use_case": "inference API", "switched_from": "", "story": "Previewing GPT\u20115.6 Sol: a next-generation model"}, {"id": 801729697, "item_id": 50108106, "url": "https://news.ycombinator.com/item?id=50108106", "by": "eknkc", "date": "2026-10-15", "stance": "neutral", "quote": "But is really fast (fireworks.ai).", "excerpt": "Feels dumber than GLM 5.3 (only open model I use extensively). But is really fast (fireworks.ai). Only checked a single, large prompt that already ran today on claude and glm though..", "context": "", "use_case": "model inference hosting", "switched_from": "", "story": "Muse Spark Open"}, {"id": 790988640, "item_id": 49436790, "url": "https://news.ycombinator.com/item?id=49436790", "by": "mark_l_watson", "date": "2026-08-25", "stance": "recommend", "quote": "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", "excerpt": "I run a lot of local models (I am always experimenting) on my 32G M2-Pro MacMini - I would love to upgrade. The financial aspects don\u2019t work however: I can learn and experiment with what I have for local models, and 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 and much faster than any reasonable home rig. Hybrid \u2018small local\u2019", "context": "runs local models on a Mac mini", "use_case": "open model inference", "switched_from": "", "story": "Apple introduces M6 and M5 Ultra"}, {"id": 790685682, "item_id": 49417855, "url": "https://news.ycombinator.com/item?id=49417855", "by": "dkersten", "date": "2026-08-24", "stance": "recommend", "quote": "I\u2019ve mostly settled on using a mixture of open weights models through Together.ai and Fireworks.ai", "excerpt": "I stopped using Claude because of this BS. If I pay for a service, I want to know what I\u2019m paying for, I want it to be predictable. Anthropic have been anything but. Flip flopping on model availability, model access behind an opaque filter, their past behaviour of model degradation as they prepared their next model\u2026 these are not signs of a reliable service. I\u2019ve mostly settled on using a mixture of open weights mode", "context": "", "use_case": "hosted open-weights models", "switched_from": "Claude", "story": "Anthropic's best AI model struggles to attract users as cheaper tools thrive"}]}, {"text": ", with cheap cached tokens and US-based zero data retention.", "cites": [{"id": 787562307, "item_id": 49222644, "url": "https://news.ycombinator.com/item?id=49222644", "by": "vijucat", "date": "2026-08-08", "stance": "recommend", "quote": "On Fireworks AI, for example, if it hits the cache, you pay only 20%.", "excerpt": "Caching makes a huge difference to cost. On Fireworks AI, for example, if it hits the cache, you pay only 20%. And uncached is just $0.14/M tokens for DSV4-0731! I get entire re-architecture projects (with new tests and documentation) done for mere dollars. DSV4-0731 is a daily driver for me. But note that you have to use Cline (or other harness) if using vscode. I was shocked at how poor the recent versions of GitHu", "context": "", "use_case": "hosted inference", "switched_from": "", "story": "DeepSeek V4 Flash 0731"}, {"id": 796663744, "item_id": 49791484, "url": "https://news.ycombinator.com/item?id=49791484", "by": "drewnick", "date": "2026-09-21", "stance": "recommend", "quote": "I use it on fireworks which is US/ZDR and pretty reliable.", "excerpt": "I use it on fireworks which is US/ZDR and pretty reliable. We run a few hundred million tokens/day through it for dollars. Many are cached, which is super duper cheap.", "context": "runs a few hundred million tokens/day", "use_case": "inference at hundreds of millions tokens/day", "switched_from": "", "story": "Grok 4.7"}, {"id": 798185424, "item_id": 49886589, "url": "https://news.ycombinator.com/item?id=49886589", "by": "verdverm", "date": "2026-09-29", "stance": "recommend", "quote": "we use Fireworks.ai, and American company with ZDR", "excerpt": "yup, we use Fireworks.ai, and American company with ZDR Their new Ember-1 model is pretty good, fine-tune of Kimi3 with way less thinking Does this make it an American model or is it still Chinese? Does it matter? https://fireworks.ai/blog/ember-1", "context": "company use with zero data retention", "use_case": "model inference hosting", "switched_from": "", "story": "Sonnet 5.5"}]}, {"text": " Complaints are scattered, with recent ones about slower new-model availability and deprecated models", "cites": [{"id": 802508800, "item_id": 50156800, "url": "https://news.ycombinator.com/item?id=50156800", "by": "verdverm", "date": "2026-10-19", "stance": "warn", "quote": "Fireworks is not keeping up with new models the last few months.", "excerpt": "Unfortunately, Fireworks is not keeping up with new models the last few months. When we selected them as a vendor, they were crushing it, new models were available in 1-2 days. Now they want us to \"request\" them in Discord, instead of making them all available. I use OpenCode at home, this is one of the reasons. Their Ember-1 model (Kimi K3 tuned to think less) hallucinates to high hell, I have no trust of it, nowher", "context": "", "use_case": "model inference provider", "switched_from": "", "story": "MiniMax M3.1 Flash"}, {"id": 785183600, "item_id": 49073975, "url": "https://news.ycombinator.com/item?id=49073975", "by": "lmf4lol", "date": "2026-07-27", "stance": "recommend", "quote": "I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform", "excerpt": "I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform for our users. Just a shame they deprecated the on-demand flux models :( Where do I get my fix for image gen now?", "context": "runs a platform with users", "use_case": "hosting Kimi-K3 inference", "switched_from": "", "story": "Kimi-K3 Releases on HuggingFace 6/27"}]}, {"text": ", plus a billing and support dispute and occasional reliability issues.", "cites": [{"id": 802198640, "item_id": 50137415, "url": "https://news.ycombinator.com/item?id=50137415", "by": "browningstreet", "date": "2026-10-18", "stance": "warn", "quote": "I'm off Fireworks.", "excerpt": "No Fireworks for me. They invited me to test Firepass, screwed up the onboarding, charged me money (when the account was supposed to be gratis) and when I asked them to fix that after I used up my $$, they said they closed the program. After that, no emails were responded to, and we were on a person-to-person basis. Not sure what their deal is, but I'm off Fireworks.", "context": "", "use_case": "model inference provider", "switched_from": "", "story": "DeepSeek-v4.1-Flash-1017"}, {"id": 785814241, "item_id": 49113390, "url": "https://news.ycombinator.com/item?id=49113390", "by": "__jl__", "date": "2026-07-30", "stance": "recommend", "quote": "Fireworks is a great provider but we still ran into issues here and there", "excerpt": "Didn't expect that. Luna pricing is crazy now. I don't think there is anything on the market that competes at this price-performance point. For our production app, OpenAI clearly is the best provider now. Their API is very reliable and has many nice features. The price-performance of the model lineup is incredible. We used open weights model via Fireworks for a long time (e.g. Kimi K2.5). Fireworks is a great provide", "context": "runs a production app", "use_case": "hosting open weights models", "switched_from": "", "story": "Advancing the price-performance frontier with GPT\u20115.6"}]}], "praise": [{"label": "Speed", "summary": "Commenters describe inference as really fast, without rate limits in daily use.", "quotes": [{"id": 781566001, "item_id": 48847875, "url": "https://news.ycombinator.com/item?id=48847875", "by": "pimeys", "date": "2026-07-09", "stance": "recommend", "quote": "I use it all the time through Fireworks.", "excerpt": "I use it all the time through Fireworks. The normal version when I pay it myself and the fast one when company pays. It's really fast and I never get rate limited with my daily use.", "context": "", "use_case": "inference hosting", "switched_from": "", "story": "Introducing Muse Spark 1.1"}, {"id": 779098850, "item_id": 48693678, "url": "https://news.ycombinator.com/item?id=48693678", "by": "mark_l_watson", "date": "2026-06-26", "stance": "recommend", "quote": "with the Fireworks.ai API and properly configured OpenCode to be very good", "excerpt": "Your experience with DeepSeek v4 Flash differs from mine: while I usually use DeepSeek v4 Pro (that is also inexpensive), I find using DeepSeek v4 Flash with the Fireworks.ai API and properly configured OpenCode to be very good for routine work, and it is pleasantly very fast. Admittedly I use DeepSeek v4 Pro for difficult problems. I encourage people to at least once a month to do a quick evaluation with their own p", "context": "", "use_case": "inference API", "switched_from": "", "story": "Previewing GPT\u20115.6 Sol: a next-generation model"}, {"id": 801729697, "item_id": 50108106, "url": "https://news.ycombinator.com/item?id=50108106", "by": "eknkc", "date": "2026-10-15", "stance": "neutral", "quote": "But is really fast (fireworks.ai).", "excerpt": "Feels dumber than GLM 5.3 (only open model I use extensively). But is really fast (fireworks.ai). Only checked a single, large prompt that already ran today on claude and glm though..", "context": "", "use_case": "model inference hosting", "switched_from": "", "story": "Muse Spark Open"}]}, {"label": "Low cost", "summary": "Users report small bills, cheap per-token pricing, and cache hits billed at 20%.", "quotes": [{"id": 787562307, "item_id": 49222644, "url": "https://news.ycombinator.com/item?id=49222644", "by": "vijucat", "date": "2026-08-08", "stance": "recommend", "quote": "On Fireworks AI, for example, if it hits the cache, you pay only 20%.", "excerpt": "Caching makes a huge difference to cost. On Fireworks AI, for example, if it hits the cache, you pay only 20%. And uncached is just $0.14/M tokens for DSV4-0731! I get entire re-architecture projects (with new tests and documentation) done for mere dollars. DSV4-0731 is a daily driver for me. But note that you have to use Cline (or other harness) if using vscode. I was shocked at how poor the recent versions of GitHu", "context": "", "use_case": "hosted inference", "switched_from": "", "story": "DeepSeek V4 Flash 0731"}, {"id": 790988640, "item_id": 49436790, "url": "https://news.ycombinator.com/item?id=49436790", "by": "mark_l_watson", "date": "2026-08-25", "stance": "recommend", "quote": "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", "excerpt": "I run a lot of local models (I am always experimenting) on my 32G M2-Pro MacMini - I would love to upgrade. The financial aspects don\u2019t work however: I can learn and experiment with what I have for local models, and 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 and much faster than any reasonable home rig. Hybrid \u2018small local\u2019", "context": "runs local models on a Mac mini", "use_case": "open model inference", "switched_from": "", "story": "Apple introduces M6 and M5 Ultra"}, {"id": 796663744, "item_id": 49791484, "url": "https://news.ycombinator.com/item?id=49791484", "by": "drewnick", "date": "2026-09-21", "stance": "recommend", "quote": "I use it on fireworks which is US/ZDR and pretty reliable.", "excerpt": "I use it on fireworks which is US/ZDR and pretty reliable. We run a few hundred million tokens/day through it for dollars. Many are cached, which is super duper cheap.", "context": "runs a few hundred million tokens/day", "use_case": "inference at hundreds of millions tokens/day", "switched_from": "", "story": "Grok 4.7"}]}, {"label": "Open-weights access", "summary": "Used as a host for open models like Kimi, GLM and DeepSeek, sometimes alongside Together.ai.", "quotes": [{"id": 790685682, "item_id": 49417855, "url": "https://news.ycombinator.com/item?id=49417855", "by": "dkersten", "date": "2026-08-24", "stance": "recommend", "quote": "I\u2019ve mostly settled on using a mixture of open weights models through Together.ai and Fireworks.ai", "excerpt": "I stopped using Claude because of this BS. If I pay for a service, I want to know what I\u2019m paying for, I want it to be predictable. Anthropic have been anything but. Flip flopping on model availability, model access behind an opaque filter, their past behaviour of model degradation as they prepared their next model\u2026 these are not signs of a reliable service. I\u2019ve mostly settled on using a mixture of open weights mode", "context": "", "use_case": "hosted open-weights models", "switched_from": "Claude", "story": "Anthropic's best AI model struggles to attract users as cheaper tools thrive"}, {"id": 779393985, "item_id": 48712124, "url": "https://news.ycombinator.com/item?id=48712124", "by": "pimeys", "date": "2026-06-28", "stance": "neutral", "quote": "I used it unquantized through Fireworks", "excerpt": "I have taken another look on these open models after the fiasco of Fable and GPT 5.6 this weekend and... GLM-5.2 truly is a good workhorse model for daily programming. I consider myself a heavy user of LLMs and a seasoned developer. A typical session for me with GPT is usually over a hundred dollars... This weekend I programmed a matrix bot with encryption and a Rust agent with some tools. Because I need one and Open", "context": "", "use_case": "GLM inference provider", "switched_from": "", "story": "Semgrep: GLM 5.2 beats Claude in our Cyber Benchmarks"}, {"id": 774708849, "item_id": 48419303, "url": "https://news.ycombinator.com/item?id=48419303", "by": "heldrida", "date": "2026-06-05", "stance": "neutral", "quote": "Opencode+Fireworks+Kimi2.5 and git worktrees mainly", "excerpt": "Opencode+Fireworks+Kimi2.5 and git worktrees mainly", "context": "", "use_case": "model inference provider", "switched_from": "", "story": "Ask HN: What is your (AI) dev tech stack / workflow? (June 2026)"}]}, {"label": "Production use", "summary": "Several commenters run production apps or user-facing platforms on it.", "quotes": [{"id": 794510256, "item_id": 49656891, "url": "https://news.ycombinator.com/item?id=49656891", "by": "ofisboy", "date": "2026-09-11", "stance": "neutral", "quote": "I use Fireworks for a production app.", "excerpt": "I use Fireworks for a production app. Much lower volumes though. My only gripe is their serverless offering for reranking has only 1 model which is Qwen 8b and it's quite expensive.", "context": "low-volume production app", "use_case": "production inference", "switched_from": "", "story": "So you want to use OpenRouter?"}, {"id": 785183600, "item_id": 49073975, "url": "https://news.ycombinator.com/item?id=49073975", "by": "lmf4lol", "date": "2026-07-27", "stance": "recommend", "quote": "I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform", "excerpt": "I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform for our users. Just a shame they deprecated the on-demand flux models :( Where do I get my fix for image gen now?", "context": "runs a platform with users", "use_case": "hosting Kimi-K3 inference", "switched_from": "", "story": "Kimi-K3 Releases on HuggingFace 6/27"}, {"id": 785814241, "item_id": 49113390, "url": "https://news.ycombinator.com/item?id=49113390", "by": "__jl__", "date": "2026-07-30", "stance": "recommend", "quote": "Fireworks is a great provider but we still ran into issues here and there", "excerpt": "Didn't expect that. Luna pricing is crazy now. I don't think there is anything on the market that competes at this price-performance point. For our production app, OpenAI clearly is the best provider now. Their API is very reliable and has many nice features. The price-performance of the model lineup is incredible. We used open weights model via Fireworks for a long time (e.g. Kimi K2.5). Fireworks is a great provide", "context": "runs a production app", "use_case": "hosting open weights models", "switched_from": "", "story": "Advancing the price-performance frontier with GPT\u20115.6"}]}, {"label": "Data retention", "summary": "Chosen as a US provider offering zero data retention.", "quotes": [{"id": 798185424, "item_id": 49886589, "url": "https://news.ycombinator.com/item?id=49886589", "by": "verdverm", "date": "2026-09-29", "stance": "recommend", "quote": "we use Fireworks.ai, and American company with ZDR", "excerpt": "yup, we use Fireworks.ai, and American company with ZDR Their new Ember-1 model is pretty good, fine-tune of Kimi3 with way less thinking Does this make it an American model or is it still Chinese? Does it matter? https://fireworks.ai/blog/ember-1", "context": "company use with zero data retention", "use_case": "model inference hosting", "switched_from": "", "story": "Sonnet 5.5"}, {"id": 796663744, "item_id": 49791484, "url": "https://news.ycombinator.com/item?id=49791484", "by": "drewnick", "date": "2026-09-21", "stance": "recommend", "quote": "I use it on fireworks which is US/ZDR and pretty reliable.", "excerpt": "I use it on fireworks which is US/ZDR and pretty reliable. We run a few hundred million tokens/day through it for dollars. Many are cached, which is super duper cheap.", "context": "runs a few hundred million tokens/day", "use_case": "inference at hundreds of millions tokens/day", "switched_from": "", "story": "Grok 4.7"}]}], "caveats": [{"label": "New model lag", "summary": "One user says new models now arrive slowly and must be requested on Discord.", "quotes": [{"id": 802508800, "item_id": 50156800, "url": "https://news.ycombinator.com/item?id=50156800", "by": "verdverm", "date": "2026-10-19", "stance": "warn", "quote": "Fireworks is not keeping up with new models the last few months.", "excerpt": "Unfortunately, Fireworks is not keeping up with new models the last few months. When we selected them as a vendor, they were crushing it, new models were available in 1-2 days. Now they want us to \"request\" them in Discord, instead of making them all available. I use OpenCode at home, this is one of the reasons. Their Ember-1 model (Kimi K3 tuned to think less) hallucinates to high hell, I have no trust of it, nowher", "context": "", "use_case": "model inference provider", "switched_from": "", "story": "MiniMax M3.1 Flash"}]}, {"label": "Model deprecations", "summary": "Users note deprecated on-demand Flux models and a single, expensive reranking model.", "quotes": [{"id": 785183600, "item_id": 49073975, "url": "https://news.ycombinator.com/item?id=49073975", "by": "lmf4lol", "date": "2026-07-27", "stance": "recommend", "quote": "I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform", "excerpt": "I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform for our users. Just a shame they deprecated the on-demand flux models :( Where do I get my fix for image gen now?", "context": "runs a platform with users", "use_case": "hosting Kimi-K3 inference", "switched_from": "", "story": "Kimi-K3 Releases on HuggingFace 6/27"}, {"id": 794510256, "item_id": 49656891, "url": "https://news.ycombinator.com/item?id=49656891", "by": "ofisboy", "date": "2026-09-11", "stance": "neutral", "quote": "I use Fireworks for a production app.", "excerpt": "I use Fireworks for a production app. Much lower volumes though. My only gripe is their serverless offering for reranking has only 1 model which is Qwen 8b and it's quite expensive.", "context": "low-volume production app", "use_case": "production inference", "switched_from": "", "story": "So you want to use OpenRouter?"}]}, {"label": "Billing and support", "summary": "One commenter was charged during a supposedly free trial and got no email responses.", "quotes": [{"id": 802198640, "item_id": 50137415, "url": "https://news.ycombinator.com/item?id=50137415", "by": "browningstreet", "date": "2026-10-18", "stance": "warn", "quote": "I'm off Fireworks.", "excerpt": "No Fireworks for me. They invited me to test Firepass, screwed up the onboarding, charged me money (when the account was supposed to be gratis) and when I asked them to fix that after I used up my $$, they said they closed the program. After that, no emails were responded to, and we were on a person-to-person basis. Not sure what their deal is, but I'm off Fireworks.", "context": "", "use_case": "model inference provider", "switched_from": "", "story": "DeepSeek-v4.1-Flash-1017"}]}, {"label": "Occasional issues", "summary": "A production user hit issues here and there and moved to OpenAI.", "quotes": [{"id": 785814241, "item_id": 49113390, "url": "https://news.ycombinator.com/item?id=49113390", "by": "__jl__", "date": "2026-07-30", "stance": "recommend", "quote": "Fireworks is a great provider but we still ran into issues here and there", "excerpt": "Didn't expect that. Luna pricing is crazy now. I don't think there is anything on the market that competes at this price-performance point. For our production app, OpenAI clearly is the best provider now. Their API is very reliable and has many nice features. The price-performance of the model lineup is incredible. We used open weights model via Fireworks for a long time (e.g. Kimi K2.5). Fireworks is a great provide", "context": "runs a production app", "use_case": "hosting open weights models", "switched_from": "", "story": "Advancing the price-performance frontier with GPT\u20115.6"}]}, {"label": "Ember-1 quality", "summary": "One user says the Ember-1 model hallucinates heavily.", "quotes": [{"id": 802508800, "item_id": 50156800, "url": "https://news.ycombinator.com/item?id=50156800", "by": "verdverm", "date": "2026-10-19", "stance": "warn", "quote": "Fireworks is not keeping up with new models the last few months.", "excerpt": "Unfortunately, Fireworks is not keeping up with new models the last few months. When we selected them as a vendor, they were crushing it, new models were available in 1-2 days. Now they want us to \"request\" them in Discord, instead of making them all available. I use OpenCode at home, this is one of the reasons. Their Ember-1 model (Kimi K3 tuned to think less) hallucinates to high hell, I have no trust of it, nowher", "context": "", "use_case": "model inference provider", "switched_from": "", "story": "MiniMax M3.1 Flash"}]}], "who": [{"by": "punkbit", "context": "Product Engineer in London", "url": "https://news.ycombinator.com/item?id=49922786", "date": "2026-10-01"}, {"by": "verdverm", "context": "company use with zero data retention", "url": "https://news.ycombinator.com/item?id=49886589", "date": "2026-09-29"}, {"by": "drewnick", "context": "runs a few hundred million tokens/day", "url": "https://news.ycombinator.com/item?id=49791484", "date": "2026-09-21"}, {"by": "ofisboy", "context": "low-volume production app", "url": "https://news.ycombinator.com/item?id=49656891", "date": "2026-09-11"}, {"by": "mark_l_watson", "context": "runs local models on a Mac mini", "url": "https://news.ycombinator.com/item?id=49436790", "date": "2026-08-25"}, {"by": "__jl__", "context": "runs a production app", "url": "https://news.ycombinator.com/item?id=49113390", "date": "2026-07-30"}, {"by": "lmf4lol", "context": "runs a platform with users", "url": "https://news.ycombinator.com/item?id=49073975", "date": "2026-07-27"}], "tags": ["for production apps", "for coding agents", "for open-weights models", "zero data retention"], "use_cases": ["model inference provider", "model inference hosting", "model inference", "inference at hundreds of millions tokens/day", "production inference", "open model inference"], "compared": [{"slug": "deepseek-v4-flash", "name": "DeepSeek V4 Flash", "co": 6, "switched_from": 0, "switched_to": 0}, {"slug": "kimi", "name": "Kimi", "co": 5, "switched_from": 0, "switched_to": 0}, {"slug": "claude", "name": "Claude", "co": 3, "switched_from": 1, "switched_to": 0}, {"slug": "opencode", "name": "OpenCode", "co": 5, "switched_from": 0, "switched_to": 0}, {"slug": "glm", "name": "GLM", "co": 3, "switched_from": 0, "switched_to": 0}, {"slug": "deepseek", "name": "DeepSeek", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "claude-code", "name": "Claude Code", "co": 2, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=50154540", "title": "MiniMax M3.1 Flash", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=50129045", "title": "DeepSeek-v4.1-Flash-1017", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=50107379", "title": "Muse Spark Open", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49922568", "title": "Ask HN: Who wants to be hired? (October 2026)", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49881850", "title": "Sonnet 5.5", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49788838", "title": "Grok 4.7", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {}, "2026-02": {}, "2026-03": {}, "2026-04": {}, "2026-05": {}, "2026-06": {"neutral": 3}, "2026-07": {"recommend": 3, "neutral": 1}, "2026-08": {"recommend": 3, "neutral": 1}, "2026-09": {"neutral": 1, "recommend": 2}, "2026-10": {"neutral": 2, "warn": 2}}, "quotes": [{"id": 802508800, "item_id": 50156800, "url": "https://news.ycombinator.com/item?id=50156800", "by": "verdverm", "date": "2026-10-19", "stance": "warn", "quote": "Fireworks is not keeping up with new models the last few months.", "excerpt": "Unfortunately, Fireworks is not keeping up with new models the last few months. When we selected them as a vendor, they were crushing it, new models were available in 1-2 days. Now they want us to \"request\" them in Discord, instead of making them all available. I use OpenCode at home, this is one of the reasons. Their Ember-1 model (Kimi K3 tuned to think less) hallucinates to high hell, I have no trust of it, nowher", "context": "", "use_case": "model inference provider", "switched_from": "", "story": "MiniMax M3.1 Flash"}, {"id": 802198640, "item_id": 50137415, "url": "https://news.ycombinator.com/item?id=50137415", "by": "browningstreet", "date": "2026-10-18", "stance": "warn", "quote": "I'm off Fireworks.", "excerpt": "No Fireworks for me. They invited me to test Firepass, screwed up the onboarding, charged me money (when the account was supposed to be gratis) and when I asked them to fix that after I used up my $$, they said they closed the program. After that, no emails were responded to, and we were on a person-to-person basis. Not sure what their deal is, but I'm off Fireworks.", "context": "", "use_case": "model inference provider", "switched_from": "", "story": "DeepSeek-v4.1-Flash-1017"}, {"id": 801729697, "item_id": 50108106, "url": "https://news.ycombinator.com/item?id=50108106", "by": "eknkc", "date": "2026-10-15", "stance": "neutral", "quote": "But is really fast (fireworks.ai).", "excerpt": "Feels dumber than GLM 5.3 (only open model I use extensively). But is really fast (fireworks.ai). Only checked a single, large prompt that already ran today on claude and glm though..", "context": "", "use_case": "model inference hosting", "switched_from": "", "story": "Muse Spark Open"}, {"id": 798764578, "item_id": 49922786, "url": "https://news.ycombinator.com/item?id=49922786", "by": "punkbit", "date": "2026-10-01", "stance": "neutral", "quote": "To iterate quickly, I rely on an inference provider such as Fireworks.", "excerpt": "Location: London (UK) Remote: Yes (Remote/Hybrid depending on location) Willing to relocate: No Technologies: Primarily Typescript, React, Bash, GitHub Actions R\u00e9sum\u00e9/CV: https://punkbit.com/docs/cv-punkbit-helder-oliveira.202607131958.pdf Email: info (at) punkbit (dot) com I'm a Product Engineer and a creative based in London. I\u2019ve built full-stack applications across web, mobile, blockchain and embedded platforms. ", "context": "Product Engineer in London", "use_case": "model inference", "switched_from": "", "story": "Ask HN: Who wants to be hired? (October 2026)"}, {"id": 798185424, "item_id": 49886589, "url": "https://news.ycombinator.com/item?id=49886589", "by": "verdverm", "date": "2026-09-29", "stance": "recommend", "quote": "we use Fireworks.ai, and American company with ZDR", "excerpt": "yup, we use Fireworks.ai, and American company with ZDR Their new Ember-1 model is pretty good, fine-tune of Kimi3 with way less thinking Does this make it an American model or is it still Chinese? Does it matter? https://fireworks.ai/blog/ember-1", "context": "company use with zero data retention", "use_case": "model inference hosting", "switched_from": "", "story": "Sonnet 5.5"}, {"id": 796663744, "item_id": 49791484, "url": "https://news.ycombinator.com/item?id=49791484", "by": "drewnick", "date": "2026-09-21", "stance": "recommend", "quote": "I use it on fireworks which is US/ZDR and pretty reliable.", "excerpt": "I use it on fireworks which is US/ZDR and pretty reliable. We run a few hundred million tokens/day through it for dollars. Many are cached, which is super duper cheap.", "context": "runs a few hundred million tokens/day", "use_case": "inference at hundreds of millions tokens/day", "switched_from": "", "story": "Grok 4.7"}, {"id": 794510256, "item_id": 49656891, "url": "https://news.ycombinator.com/item?id=49656891", "by": "ofisboy", "date": "2026-09-11", "stance": "neutral", "quote": "I use Fireworks for a production app.", "excerpt": "I use Fireworks for a production app. Much lower volumes though. My only gripe is their serverless offering for reranking has only 1 model which is Qwen 8b and it's quite expensive.", "context": "low-volume production app", "use_case": "production inference", "switched_from": "", "story": "So you want to use OpenRouter?"}, {"id": 790988640, "item_id": 49436790, "url": "https://news.ycombinator.com/item?id=49436790", "by": "mark_l_watson", "date": "2026-08-25", "stance": "recommend", "quote": "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", "excerpt": "I run a lot of local models (I am always experimenting) on my 32G M2-Pro MacMini - I would love to upgrade. The financial aspects don\u2019t work however: I can learn and experiment with what I have for local models, and 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 and much faster than any reasonable home rig. Hybrid \u2018small local\u2019", "context": "runs local models on a Mac mini", "use_case": "open model inference", "switched_from": "", "story": "Apple introduces M6 and M5 Ultra"}, {"id": 790685682, "item_id": 49417855, "url": "https://news.ycombinator.com/item?id=49417855", "by": "dkersten", "date": "2026-08-24", "stance": "recommend", "quote": "I\u2019ve mostly settled on using a mixture of open weights models through Together.ai and Fireworks.ai", "excerpt": "I stopped using Claude because of this BS. If I pay for a service, I want to know what I\u2019m paying for, I want it to be predictable. Anthropic have been anything but. Flip flopping on model availability, model access behind an opaque filter, their past behaviour of model degradation as they prepared their next model\u2026 these are not signs of a reliable service. I\u2019ve mostly settled on using a mixture of open weights mode", "context": "", "use_case": "hosted open-weights models", "switched_from": "Claude", "story": "Anthropic's best AI model struggles to attract users as cheaper tools thrive"}, {"id": 789272724, "item_id": 49329545, "url": "https://news.ycombinator.com/item?id=49329545", "by": "squidbeak", "date": "2026-08-17", "stance": "recommend", "quote": "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.", "excerpt": "This isn't true. 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. GIMP's UX is just bad. I loathe photoshop and desperately want to like GIMP, but each time I make the attempt the friction is too much and the workflows nonsensical. I used it a few weeks ago to place two images side by side in a larger ", "context": "", "use_case": "image editing", "switched_from": "", "story": "GIMP Development Update"}, {"id": 787742050, "item_id": 49233878, "url": "https://news.ycombinator.com/item?id=49233878", "by": "mark_l_watson", "date": "2026-08-09", "stance": "neutral", "quote": "hardwired to use deepseek-v4-flash-0731 via FireWorks.ai", "excerpt": "I am retired, so this may not be of general interest: I am working on my fourth AI coding agent, this one written in Racket Scheme: https://github.com/mark-watson/racket-coding-agent I don't necessarily suggest that anyone besides myself use it. I open-sourced it to perhaps give other people ideas, but it is tailored to just my own needs (e.g., hardwired to use deepseek-v4-flash-0731 via FireWorks.ai; supports just t", "context": "", "use_case": "model hosting", "switched_from": "", "story": "Ask HN: What are you working on? (August 2026)"}, {"id": 787562307, "item_id": 49222644, "url": "https://news.ycombinator.com/item?id=49222644", "by": "vijucat", "date": "2026-08-08", "stance": "recommend", "quote": "On Fireworks AI, for example, if it hits the cache, you pay only 20%.", "excerpt": "Caching makes a huge difference to cost. On Fireworks AI, for example, if it hits the cache, you pay only 20%. And uncached is just $0.14/M tokens for DSV4-0731! I get entire re-architecture projects (with new tests and documentation) done for mere dollars. DSV4-0731 is a daily driver for me. But note that you have to use Cline (or other harness) if using vscode. I was shocked at how poor the recent versions of GitHu", "context": "", "use_case": "hosted inference", "switched_from": "", "story": "DeepSeek V4 Flash 0731"}, {"id": 785949026, "item_id": 49121814, "url": "https://news.ycombinator.com/item?id=49121814", "by": "mark_l_watson", "date": "2026-07-31", "stance": "neutral", "quote": "I buy tokens mostly from FireWorks.ai in the US", "excerpt": "This is my experience also: DeepSeek v4 flash is good enough for most of my work and I like the fast response times. I buy tokens mostly from FireWorks.ai in the US, but I also prepaid for a large chunk of tokens directoy with DeepSeek. I use OpenCode mostly (uses fewer tokens than Claude Code) and I am looking forward to the release of DeepSeek\u2019s own coding harness.", "context": "", "use_case": "buying inference tokens", "switched_from": "", "story": "DeepSeek-V4-Flash Update"}, {"id": 785814241, "item_id": 49113390, "url": "https://news.ycombinator.com/item?id=49113390", "by": "__jl__", "date": "2026-07-30", "stance": "recommend", "quote": "Fireworks is a great provider but we still ran into issues here and there", "excerpt": "Didn't expect that. Luna pricing is crazy now. I don't think there is anything on the market that competes at this price-performance point. For our production app, OpenAI clearly is the best provider now. Their API is very reliable and has many nice features. The price-performance of the model lineup is incredible. We used open weights model via Fireworks for a long time (e.g. Kimi K2.5). Fireworks is a great provide", "context": "runs a production app", "use_case": "hosting open weights models", "switched_from": "", "story": "Advancing the price-performance frontier with GPT\u20115.6"}, {"id": 785183600, "item_id": 49073975, "url": "https://news.ycombinator.com/item?id=49073975", "by": "lmf4lol", "date": "2026-07-27", "stance": "recommend", "quote": "I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform", "excerpt": "I love fireworks.ai! They launched it couple of hours ago and we have it now already live on our platform for our users. Just a shame they deprecated the on-demand flux models :( Where do I get my fix for image gen now?", "context": "runs a platform with users", "use_case": "hosting Kimi-K3 inference", "switched_from": "", "story": "Kimi-K3 Releases on HuggingFace 6/27"}, {"id": 781566001, "item_id": 48847875, "url": "https://news.ycombinator.com/item?id=48847875", "by": "pimeys", "date": "2026-07-09", "stance": "recommend", "quote": "I use it all the time through Fireworks.", "excerpt": "I use it all the time through Fireworks. The normal version when I pay it myself and the fast one when company pays. It's really fast and I never get rate limited with my daily use.", "context": "", "use_case": "inference hosting", "switched_from": "", "story": "Introducing Muse Spark 1.1"}, {"id": 779393985, "item_id": 48712124, "url": "https://news.ycombinator.com/item?id=48712124", "by": "pimeys", "date": "2026-06-28", "stance": "neutral", "quote": "I used it unquantized through Fireworks", "excerpt": "I have taken another look on these open models after the fiasco of Fable and GPT 5.6 this weekend and... GLM-5.2 truly is a good workhorse model for daily programming. I consider myself a heavy user of LLMs and a seasoned developer. A typical session for me with GPT is usually over a hundred dollars... This weekend I programmed a matrix bot with encryption and a Rust agent with some tools. Because I need one and Open", "context": "", "use_case": "GLM inference provider", "switched_from": "", "story": "Semgrep: GLM 5.2 beats Claude in our Cyber Benchmarks"}, {"id": 779098850, "item_id": 48693678, "url": "https://news.ycombinator.com/item?id=48693678", "by": "mark_l_watson", "date": "2026-06-26", "stance": "recommend", "quote": "with the Fireworks.ai API and properly configured OpenCode to be very good", "excerpt": "Your experience with DeepSeek v4 Flash differs from mine: while I usually use DeepSeek v4 Pro (that is also inexpensive), I find using DeepSeek v4 Flash with the Fireworks.ai API and properly configured OpenCode to be very good for routine work, and it is pleasantly very fast. Admittedly I use DeepSeek v4 Pro for difficult problems. I encourage people to at least once a month to do a quick evaluation with their own p", "context": "", "use_case": "inference API", "switched_from": "", "story": "Previewing GPT\u20115.6 Sol: a next-generation model"}, {"id": 778072657, "item_id": 48629541, "url": "https://news.ycombinator.com/item?id=48629541", "by": "mark_l_watson", "date": "2026-06-22", "stance": "neutral", "quote": "I use the US company Fireworks.ai", "excerpt": "But, it is not all about cost: models like DeepSeek v4 flash (I use the US company Fireworks.ai and also buy tokens directly from DeepSeek) is very fast, very low latency while working. Would you want to use a text editor that updates the screen very slowly? Kind of the same thing for using agentic systems as coding assistants: don\u2019t want a \u2018sluggish\u2019 experience.", "context": "", "use_case": "model inference provider", "switched_from": "", "story": "Sakana Fugu"}, {"id": 774708849, "item_id": 48419303, "url": "https://news.ycombinator.com/item?id=48419303", "by": "heldrida", "date": "2026-06-05", "stance": "neutral", "quote": "Opencode+Fireworks+Kimi2.5 and git worktrees mainly", "excerpt": "Opencode+Fireworks+Kimi2.5 and git worktrees mainly", "context": "", "use_case": "model inference provider", "switched_from": "", "story": "Ask HN: What is your (AI) dev tech stack / workflow? (June 2026)"}]}