{"entity":"fireworks-ai","id":"ent_fireworks-ai","name":"Fireworks AI","wikidata":"Q140774682","page":"https://asof.is/record/fireworks-ai","as_of":"2026-10-28","known_by":null,"since":null,"fact_pages":[{"predicate":"headquarters","page":"https://asof.is/record/fireworks-ai/headquarters","md":"https://asof.is/record/fireworks-ai/headquarters.md","json":"https://asof.is/v1/entities/fireworks-ai/headquarters"},{"predicate":"founded","page":"https://asof.is/record/fireworks-ai/founded","md":"https://asof.is/record/fireworks-ai/founded.md","json":"https://asof.is/v1/entities/fireworks-ai/founded"}],"coverage":{"entity":"fireworks-ai","name":"Fireworks AI","kind":"organization","sources":[{"source":"wikidata","item":"Q140774682","covers":"dated statements whose reference URL is the source: leadership, ownership, names, founding, headquarters; never financial figures"},{"source":"wikidata-products","covers":"products it makes, develops or publishes, as released claims, when the release date cites a reference URL; a rotating scan, a few dozen entities a run"},{"source":"feed","feed":"fireworks-blog","name":"Fireworks AI Blog","url":"https://fireworks.ai/sitemap.xml","records_on":"fireworks-ai","covers":"its posts are recorded here; each claim extracted from a post is recorded on the entity it is about"}],"holds":{"claims":46,"open":45,"retracted":0,"first":"2022-09-01","predicates":{"investment":{"claims":4,"open":4,"kind":"event"},"feature_added":{"claims":13,"open":13,"kind":"event"},"released":{"claims":9,"open":9,"kind":"event"},"available_in":{"claims":4,"open":4,"kind":"event"},"customer":{"claims":3,"open":3,"kind":"event"},"launched":{"claims":4,"open":4,"kind":"event"},"adopted":{"claims":1,"open":1,"kind":"event"},"partnership":{"claims":3,"open":3,"kind":"event"},"founded":{"claims":1,"open":1,"kind":"fact"},"acquired":{"claims":1,"open":1,"kind":"event"},"price":{"claims":1,"open":0,"kind":"fact"},"certified":{"claims":1,"open":1,"kind":"event"},"headquarters":{"claims":1,"open":1,"kind":"fact"}}},"freshness":[{"origin":"manual","label":"a one-off fix read from primary sources (pipeline/fixes)","claims":2,"last_added":"2026-10-31","last_checked":"2026-10-31","last_verified":null},{"origin":"feed","label":"claims extracted from official feeds (feeds_ingest.py)","claims":44,"last_added":"2026-10-30","last_checked":"2026-10-31","last_verified":null}],"gaps":["ceo: none public; unknown, not absent: a claim is published only once checked against its source","financials: none; not an SEC filer on the record (pipeline/sec_companies.json), so figures come only from the financials audit or two press outlets"],"group":[],"group_complete":true,"rule":"sources: what the pipeline reads for this entity, from its config (pipeline/entities.json, feeds.json, sec_companies.json, huggingface_authors.json). holds: its public claims by predicate; open is held today; kind fact is what entity_state answers by default, event is for timeline. freshness, per ingest (origin): last_added is the day it last put a new claim here (first_seen_at), last_checked the day one was last checked against its source (checked_on), last_verified the day reverify.py last confirmed one against Wikidata (value_json.verified_at). Only checked claims are public, so a fact missing here is unknown, not absent. gaps: what the record does not cover for this entity, so an empty answer is not read as news."},"claims":[{"id":"clm_864E47F4D2BD","predicate":"acquired","value":"Hathora","valid_from":"2026-03-08","valid_to":null,"published":"2026-03-08","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireworks-hathora-acquisition","tier":"primary","publisher":"Fireworks AI","published_at":"2026-03-08T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"MAR 08, 2026 Fireworks acquires Hathora to accelerate global compute orchestration Fireworks has acquired Hathora, and we’re thrilled to bring their team and technology into the Fireworks family.","url":"https://fireworks.ai/blog/fireworks-hathora-acquisition","excerpt":"Fireworks acquired Hathora, whose team built container orchestration across 14 regions and four clouds. See how it brings faster routing to your AI apps."}},{"id":"clm_3A8B30DB1673","predicate":"adopted","value":"SSD-1B model","valid_from":"2023-11-02","valid_to":null,"published":"2023-11-02","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/sdxl-ssd-1b-image-to-image-controlnet","tier":"primary","publisher":"Fireworks AI","published_at":"2023-11-02T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Try out the SSD-1B model in the Fireworks console !","url":"https://fireworks.ai/blog/sdxl-ssd-1b-image-to-image-controlnet","excerpt":"Fireworks adds Segmind's SSD-1B next to SDXL. Run image-to-image or ControlNet on either model at nine native resolutions, with PNG or JPEG output."}},{"id":"clm_F77C4757B8BE","predicate":"available_in","value":"US serverless","valid_from":"2026-10-20","valid_to":null,"published":"2026-10-20","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/meta-muse-spark-open","tier":"primary","publisher":"Fireworks AI","published_at":"2026-10-20T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Muse Spark Open on Fireworks: Meta’s open weights model, available on US serverless Muse Spark Open, an open weights model from Meta, is now available on Fireworks.","url":"https://fireworks.ai/blog/meta-muse-spark-open","excerpt":"Muse Spark Open, Meta’s open weights model, is live on Fireworks: #1 on Vals AI’s Harvey Legal Agent Benchmark, on US-only serverless with zero data retention."}},{"id":"clm_34E3E67BC648","predicate":"available_in","value":"US-only Serverless","valid_from":"2026-07-27","valid_to":null,"published":"2026-07-27","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/kimi-k3-fireworks","tier":"primary","publisher":"Fireworks AI","published_at":"2026-07-27T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Today, we’re also launching US-only Serverless endpoints, starting with Kimi K3. Over the coming days and weeks, we are adding other US-only endpoints for the most popular models on Fireworks.","url":"https://fireworks.ai/blog/kimi-k3-fireworks","excerpt":"Run Kimi K3 (2.8T params, 1M context) on US-only Serverless with Zero Data Retention by default. Train it Day-0 in the Serverless Training private preview."}},{"id":"clm_12217BD28DCA","predicate":"available_in","value":"MiniMax M3 on Fireworks","valid_from":"2026-06-12","valid_to":null,"published":"2026-06-12","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/minimax-m3","tier":"primary","publisher":"Fireworks AI","published_at":"2026-06-12T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"We are launching Day-0 support for MiniMax M3 on Fireworks.","url":"https://fireworks.ai/blog/minimax-m3","excerpt":"MiniMax M3 is live on Fireworks with Day-0 support: 500K-token context, image and video input, and agentic coding at M2.7 pricing. Start building today."}},{"id":"clm_475BB063CCC3","predicate":"available_in","value":"NVIDIA Nemotron 3 Nano","valid_from":"2025-12-15","valid_to":null,"published":"2025-12-15","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/nvidia-nemotron-3-nano","tier":"primary","publisher":"Fireworks AI","published_at":"2025-12-15T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"We’re excited to launch Day-0 support on Fireworks for the latest model in the NVIDIA Nemotron family, NVIDIA Nemotron 3 Nano, an advanced reasoning model set to fuel the next generation of AI Agents.","url":"https://fireworks.ai/blog/nvidia-nemotron-3-nano","excerpt":"Run NVIDIA Nemotron 3 Nano on Fireworks: a hybrid MoE reasoning model with 30B parameters, 3B active, and 1M context. See how it works and how to deploy it."}},{"id":"clm_C01EF30B9E9F","predicate":"certified","value":"SOC 2 Type II","valid_from":"2023-10-27","valid_to":null,"published":"2023-10-27","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireworks-soc-2-type-ii-hipaa-compliance","tier":"primary","publisher":"Fireworks AI","published_at":"2023-10-27T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Announcement OCT 27, 2023 Fireworks.ai achieves SOC 2 Type II and HIPAA compliance We are pleased to report that the Fireworks.ai inference platform is both SOC 2 Type II and HIPAA compliant.","url":"https://fireworks.ai/blog/fireworks-soc-2-type-ii-hipaa-compliance","excerpt":"Fireworks passed a third-party SOC 2 Type II audit and runs routine penetration tests. See how your healthcare team can build chatbots under HIPAA."}},{"id":"clm_216959C42018","predicate":"customer","value":"Vercel","valid_from":"2025-06-09","valid_to":null,"published":"2025-06-09","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/reinforcement-fine-tuning","tier":"primary","publisher":"Fireworks AI","published_at":"2025-06-09T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Vercel and Fireworks collaborated to train a custom code fixing model using RFT that matched the quality of closed models while running 10 - 40 times faster!","url":"https://fireworks.ai/blog/reinforcement-fine-tuning","excerpt":"Train open models with RFT: write a Python evaluator that scores outputs. See how Vercel matched closed-model quality while running 10–40x faster."}},{"id":"clm_736B9207C7BA","predicate":"customer","value":"AlliumAI","valid_from":"2024-09-25","valid_to":null,"published":"2024-09-25","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/enterprise-multimodal-model-use-cases","tier":"primary","publisher":"Fireworks AI","published_at":"2024-09-25T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"AlliumAI uses Fireworks Serverless to help retailers boost e-commerce sales by extracting and structuring product catalog data in real-time with high cost-efficiency","url":"https://fireworks.ai/blog/enterprise-multimodal-model-use-cases","excerpt":"Healthcare and insurance teams process records at 100x lower cost than GPT-4o, and AlliumAI structures product catalogs with multimodal models on Fireworks."}},{"id":"clm_CBECA92DD9C4","predicate":"customer","value":"Sourcegraph","valid_from":"2023-10-11","valid_to":null,"published":"2023-10-11","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fast-llm-inference-code-completion","tier":"primary","publisher":"Fireworks AI","published_at":"2023-10-11T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Sourcegraph’s Cody now utilizes the StarCoder model running on Fireworks for the majority of its completion in the community edition.","url":"https://fireworks.ai/blog/fast-llm-inference-code-completion","excerpt":"Sourcegraph's Cody doubled its completion acceptance rate to 30% and cut single-line latency to 1.1s with StarCoder on Fireworks. See the benchmarks."}},{"id":"clm_8DBE0164AAEE","predicate":"feature_added","value":"Zero data retention","valid_from":"2026-10-20","valid_to":null,"published":"2026-10-20","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/meta-muse-spark-open","tier":"primary","publisher":"Fireworks AI","published_at":"2026-10-20T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Muse Spark Open was trained and developed in the US, and Fireworks runs it with Zero Data Retention by default.","url":"https://fireworks.ai/blog/meta-muse-spark-open","excerpt":"Muse Spark Open, Meta’s open weights model, is live on Fireworks: #1 on Vals AI’s Harvey Legal Agent Benchmark, on US-only serverless with zero data retention."}},{"id":"clm_C08CD5B1F99B","predicate":"feature_added","value":"GLM 5.2 Fast serverless","valid_from":"2026-06-30","valid_to":null,"published":"2026-06-30","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/glm-5-2-fast-inference","tier":"primary","publisher":"Fireworks AI","published_at":"2026-06-30T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"GLM 5.2 Fast serverless inference Model release JUN 30, 2026 GLM 5.2 Fast is live on Fireworks GLM 5.2 Fast is live on Fireworks today.","url":"https://fireworks.ai/blog/glm-5-2-fast-inference","excerpt":"Run GLM 5.2 Fast on Fireworks serverless with no reserved GPUs: same weights and quality, full 1M-token context, enabled with one model ID switch."}},{"id":"clm_070F87CE648B","predicate":"feature_added","value":"Full-parameter RL tuning, private preview","valid_from":"2026-01-26","valid_to":null,"published":"2026-01-26","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/kimi-k2-5-benchmarks-api","tier":"primary","publisher":"Fireworks AI","published_at":"2026-01-26T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Additionally, we now offer a full parameter RL tuning private preview for Kimi K2.5, enabling application builders to fine tune the SOTA OSS VLM model for use cases like vibe coding and agentic workflows.","url":"https://fireworks.ai/blog/kimi-k2-5-benchmarks-api","excerpt":"Run Kimi K2.5 on Fireworks with Day-0 support at up to 200 tokens/s, plus a private preview of full-parameter RL tuning for agentic and vibe coding work."}},{"id":"clm_08DFFBA70DF8","predicate":"feature_added","value":"Batch API","valid_from":"2025-07-31","valid_to":null,"published":"2025-07-31","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/batch-inference-api","tier":"primary","publisher":"Fireworks AI","published_at":"2025-07-31T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Announcement JUL 31, 2025 Run bulk async workloads with Fireworks Batch API","url":"https://fireworks.ai/blog/batch-inference-api","excerpt":"Run large async LLM jobs on 1,000+ models with the Fireworks Batch API. Upload a JSONL dataset and get results within 24 hours at half the serverless price."}},{"id":"clm_1953F7F3B7F6","predicate":"feature_added","value":"Quantization-aware training","valid_from":"2025-06-13","valid_to":null,"published":"2025-06-13","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/supervised-fine-tuning-v2","tier":"primary","publisher":"Fireworks AI","published_at":"2025-06-13T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Quantization Aware Training To ensure optimal inference quality, both FP4 and FP8 quantization-aware training options are supported.","url":"https://fireworks.ai/blog/supervised-fine-tuning-v2","excerpt":"Supervised Fine Tuning V2 trains 2x faster than V1, supports Qwen, Llama, and DeepSeek models, and adds quantization-aware training. See what's new."}},{"id":"clm_8B6641F51FC4","predicate":"feature_added","value":"Qwen, Llama, DeepSeek SFT support","valid_from":"2025-06-13","valid_to":null,"published":"2025-06-13","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/supervised-fine-tuning-v2","tier":"primary","publisher":"Fireworks AI","published_at":"2025-06-13T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"This now encompasses the Qwen 2/2.5/3 series, Phi 4, Gemma 3, the Llama family, as well as leading open-source MoE models like Deepseek R1 and V3, including, of course, any fine-tuned variants of those models!","url":"https://fireworks.ai/blog/supervised-fine-tuning-v2","excerpt":"Supervised Fine Tuning V2 trains 2x faster than V1, supports Qwen, Llama, and DeepSeek models, and adds quantization-aware training. See what's new."}},{"id":"clm_F10A6ABF4269","predicate":"feature_added","value":"Supervised Fine Tuning V2","valid_from":"2025-06-13","valid_to":null,"published":"2025-06-13","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/supervised-fine-tuning-v2","tier":"primary","publisher":"Fireworks AI","published_at":"2025-06-13T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Announcement JUN 13, 2025 Introducing Supervised Fine Tuning V2","url":"https://fireworks.ai/blog/supervised-fine-tuning-v2","excerpt":"Supervised Fine Tuning V2 trains 2x faster than V1, supports Qwen, Llama, and DeepSeek models, and adds quantization-aware training. See what's new."}},{"id":"clm_A2994982CB54","predicate":"feature_added","value":"Document Inlining","valid_from":"2024-12-18","valid_to":null,"published":"2024-12-18","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/deepseek-v3-vision-document-inlining","tier":"primary","publisher":"Fireworks AI","published_at":"2024-12-18T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"You can now give DeepSeek V3 or any other LLM on Fireworks Model Library vision capabilities, using Fireworks’ recent feature - Document Inlining!","url":"https://fireworks.ai/blog/deepseek-v3-vision-document-inlining","excerpt":"See how Document Inlining let text-only DeepSeek V3 read images and PDFs on Fireworks by adding #transform=inline to a file URL."}},{"id":"clm_04786F1201A0","predicate":"feature_added","value":"Voice activity detection","valid_from":"2024-12-09","valid_to":null,"published":"2024-12-09","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fast-whisper-transcription-api","tier":"primary","publisher":"Fireworks AI","published_at":"2024-12-09T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Fireworks now supports audio transcription, with a fast and comprehensive audio model. Transcribe 1 hour of audio in 4 seconds and use additional features like audio-transcription alignment and voice activity detection.","url":"https://fireworks.ai/blog/fast-whisper-transcription-api","excerpt":"Transcribe 1 hour of audio in 4 seconds with Whisper v3-large on Fireworks, plus built-in translation, word-level alignment, and voice activity detection."}},{"id":"clm_2EEBB578798E","predicate":"feature_added","value":"Serverless Llama 3.1 inference","valid_from":"2024-07-23","valid_to":null,"published":"2024-07-23","recorded":"2026-10-30","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/llama-3-1-405b-api","tier":"primary","publisher":"Fireworks AI","published_at":"2024-07-23T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Llama 3.1 405B serverless inference endpoints Model release JUL 23, 2024 Introducing Llama 3.1 inference endpoints in partnership with Meta","url":"https://fireworks.ai/blog/llama-3-1-405b-api","excerpt":"Run Llama 3.1 405B, 70B, and 8B Instruct on Fireworks serverless inference, with 128K context and tool calling. Get an API key and call it from Python."}},{"id":"clm_0A9CE74A411B","predicate":"feature_added","value":"LoRA fine-tuning service","valid_from":"2024-03-08","valid_to":null,"published":"2024-03-08","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/lora-fine-tuning-service-launch","tier":"primary","publisher":"Fireworks AI","published_at":"2024-03-08T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Fine-tuning service launch Fireworks is offering an affordable LoRA fine-tuning service for 10 models, including Mixtral.","url":"https://fireworks.ai/blog/lora-fine-tuning-service-launch","excerpt":"Fireworks launched LoRA fine-tuning in March 2024 for 10 models, including Mixtral. See the launch pricing, Firectl workflow, and how to deploy tuned models."}},{"id":"clm_099658E2750E","predicate":"feature_added","value":"JSON mode","valid_from":"2024-02-20","valid_to":null,"published":"2024-02-20","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/llm-structured-output-json-grammar-mode","tier":"primary","publisher":"Fireworks AI","published_at":"2024-02-20T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Today, we’re launching of two features on all Fireworks language models: JSON mode - Always generate valid JSON according to a provided schema","url":"https://fireworks.ai/blog/llm-structured-output-json-grammar-mode","excerpt":"Learn why LLMs need structured output and how Fireworks JSON mode and grammar mode constrain responses to your JSON schema or BNF grammar. See code examples."}},{"id":"clm_B121F315DA1D","predicate":"feature_added","value":"Grammar mode","valid_from":"2024-02-20","valid_to":null,"published":"2024-02-20","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/llm-structured-output-json-grammar-mode","tier":"primary","publisher":"Fireworks AI","published_at":"2024-02-20T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Announcement FEB 20, 2024 Why do all LLMs need structured output modes? Fireworks introduces “JSON mode” and “grammar mode” for 100% adherence to provided output schemas","url":"https://fireworks.ai/blog/llm-structured-output-json-grammar-mode","excerpt":"Learn why LLMs need structured output and how Fireworks JSON mode and grammar mode constrain responses to your JSON schema or BNF grammar. See code examples."}},{"id":"clm_4393180B971E","predicate":"founded","value":"2022-09-01","valid_from":"2022-09-01","valid_to":null,"published":"2024-11-25","recorded":"2026-10-31","confidence":0.95,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://www.businessinsider.com/fireworks-ai-is-seeking-new-funding-at-around-valuation-2026-10","tier":"secondary","publisher":"Business Insider","published_at":"2026-10-20T00:00:00+00:00"},{"url":"https://latent.space/p/fireworks","tier":"secondary","publisher":"Latent Space","published_at":"2024-11-25T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Founded in 2022, Fireworks was built by veterans of Meta and Google's AI efforts.","url":"https://www.businessinsider.com/fireworks-ai-is-seeking-new-funding-at-around-valuation-2026-10"}},{"id":"clm_EFAEA82004C8","predicate":"headquarters","value":"San Mateo","valid_from":null,"undated":true,"valid_to":null,"published":"2026-07-16","recorded":"2026-10-31","confidence":0.95,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://thelogic.co/briefing/ontario-teachers-pension-plan-backs-fireworks-in-us1-5b-round-to-run-ai-for-businesses/","tier":"secondary","publisher":"The Logic","published_at":"2026-07-17T00:00:00+00:00"},{"url":"https://www.cnbc.com/2026/07/16/fireworks-nvidia-cloud-ai-startup-value.html","tier":"secondary","publisher":"cnbc.com","published_at":"2026-07-16T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Atreides Management, Index Ventures and TCV led the Series D round, which values the San Mateo, Calif.-based firm at US$17.5 billion.","url":"https://thelogic.co/briefing/ontario-teachers-pension-plan-backs-fireworks-in-us1-5b-round-to-run-ai-for-businesses/"}},{"id":"clm_A2FA0F57F753","predicate":"investment","value":"$1.505 billion Series D","valid_from":"2026-07-15","valid_to":null,"published":"2026-07-15","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireworks-series-d-funding","tier":"primary","publisher":"Fireworks AI","published_at":"2026-07-15T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Today, Fireworks announced a $1.505 billion Series D at a $17.5 billion valuation, led by Atreides Management, Index Ventures, and TCV, with participation from Evantic Capital, Lightspeed Venture Partners, Nvidia, 20VC, Bessemer Venture Partners, Menlo Ventures, and others.","url":"https://fireworks.ai/blog/fireworks-series-d-funding","excerpt":"Fireworks raised $1.505 billion at a $17.5 billion valuation, led by Atreides Management, Index Ventures, and TCV, and passed $1 billion in annualized revenue."}},{"id":"clm_E992120FDDE0","predicate":"investment","value":"$250M Series C","valid_from":"2025-10-28","valid_to":null,"published":"2025-10-28","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireworks-ai-series-c-funding","tier":"primary","publisher":"Fireworks AI","published_at":"2025-10-28T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Today, we’re announcing a $250 million Series C at a $4 billion valuation, co-led by Lightspeed Venture Partners, Index Ventures, and Evantic, with continued support from Sequoia Capital.","url":"https://fireworks.ai/blog/fireworks-ai-series-c-funding","excerpt":"Fireworks raised a $250M Series C at a $4B valuation, co-led by Lightspeed Venture Partners, Index Ventures, and Evantic. See what it funds next."}},{"id":"clm_0360A0DDC898","predicate":"investment","value":"$52M Series B","valid_from":"2024-07-11","valid_to":null,"published":"2024-07-11","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireworks-series-b-funding-sequoia","tier":"primary","publisher":"Fireworks AI","published_at":"2024-07-11T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"JUL 11, 2024 Fireworks raises $52M Series B to lead industry shift to compound AI systems","url":"https://fireworks.ai/blog/fireworks-series-b-funding-sequoia","excerpt":"Fireworks raised a $52M Series B led by Sequoia at a $552M valuation, with NVIDIA, AMD, and MongoDB Ventures joining, to build compound AI systems."}},{"id":"clm_7891BE09498B","predicate":"investment","value":"$552M valuation","valid_from":"2024-07-11","valid_to":null,"published":"2024-07-11","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireworks-series-b-funding-sequoia","tier":"primary","publisher":"Fireworks AI","published_at":"2024-07-11T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"We’re thrilled to announce our $52M Series B funding round led by Sequoia Capital, raising our valuation to $552M.","url":"https://fireworks.ai/blog/fireworks-series-b-funding-sequoia","excerpt":"Fireworks raised a $52M Series B led by Sequoia at a $552M valuation, with NVIDIA, AMD, and MongoDB Ventures joining, to build compound AI systems."}},{"id":"clm_21DF7949F4D8","predicate":"launched","value":"Voyage Multimodal 3.5","valid_from":"2026-08-05","valid_to":null,"published":"2026-08-05","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/voyage-ai-embedding-reranking-models","tier":"primary","publisher":"Fireworks AI","published_at":"2026-08-05T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-29","quote":"The full Voyage lineup now runs natively on Fireworks: the Voyage 4 family, voyage-multimodal-3.5, and rerank-2.5.","url":"https://fireworks.ai/blog/voyage-ai-embedding-reranking-models","excerpt":"Run Voyage 4 embeddings, Voyage Multimodal 3.5, and Voyage ReRank 2.5 on Fireworks, so your RAG pipeline embeds, reranks, and generates behind one API."}},{"id":"clm_7FF4823460DA","predicate":"launched","value":"Kimi K3","valid_from":"2026-07-27","valid_to":null,"published":"2026-07-27","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/kimi-k3-fireworks","tier":"primary","publisher":"Fireworks AI","published_at":"2026-07-27T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Kimi K3 is open-weight today, and available Day-0 for both inference and training on Fireworks.","url":"https://fireworks.ai/blog/kimi-k3-fireworks","excerpt":"Run Kimi K3 (2.8T params, 1M context) on US-only Serverless with Zero Data Retention by default. Train it Day-0 in the Serverless Training private preview."}},{"id":"clm_B730F1C5853E","predicate":"launched","value":"Fireworks Training","valid_from":"2026-04-06","valid_to":null,"published":"2026-04-06","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireworks-training-preview-fine-tuning","tier":"primary","publisher":"Fireworks AI","published_at":"2026-04-06T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-29","quote":"Announcement APR 06, 2026 Own your AI: Fireworks Training preview Introducing Fireworks Training in preview: an end-to-end platform for training and deploying frontier models at scale, with three surfaces for three kinds of teams.","url":"https://fireworks.ai/blog/fireworks-training-preview-fine-tuning","excerpt":"Train open models with SFT, DPO, or RL, from Qwen3 8B to 1T-parameter Kimi K2.5. Compare Training Agent, Managed Training, and the Training API."}},{"id":"clm_AEFF6FFF2CC2","predicate":"launched","value":"Virtual Cloud","valid_from":"2025-06-16","valid_to":null,"published":"2025-06-16","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireworks-virtual-cloud","tier":"primary","publisher":"Fireworks AI","published_at":"2025-06-16T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-29","quote":"That’s why today we’re excited to announce the GA of the Fireworks Virtual Cloud , a platform that abstracts away the complexity of managing GPU deployments, handling hardware failures, and scaling workloads across a global fleet.","url":"https://fireworks.ai/blog/fireworks-virtual-cloud","excerpt":"Fireworks Virtual Cloud schedules your AI workloads across regions and clouds, recovers from GPU failures automatically, and supports BYOC. See how it works."}},{"id":"clm_3E33808B8984","predicate":"partnership","value":"Voyage AI","valid_from":"2026-08-05","valid_to":null,"published":"2026-08-05","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/voyage-ai-embedding-reranking-models","tier":"primary","publisher":"Fireworks AI","published_at":"2026-08-05T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"Fireworks is the first and only dedicated inference platform Voyage AI by MongoDB has partnered with. The full Voyage lineup now runs natively on Fireworks: the Voyage 4 family, voyage-multimodal-3.5, and rerank-2.5.","url":"https://fireworks.ai/blog/voyage-ai-embedding-reranking-models","excerpt":"Run Voyage 4 embeddings, Voyage Multimodal 3.5, and Voyage ReRank 2.5 on Fireworks, so your RAG pipeline embeds, reranks, and generates behind one API."}},{"id":"clm_9CE27DF4423D","predicate":"partnership","value":"AMD","valid_from":"2025-10-20","valid_to":null,"published":"2025-10-20","recorded":"2026-10-30","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireworks-amd-instinct-gpu-partnership","tier":"primary","publisher":"Fireworks AI","published_at":"2025-10-20T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"OCT 20, 2025 Fireworks and AMD partner to power next-generation AI infrastructure on AMD Instinct™ GPUs Fireworks and AMD have entered a multi-year partnership to optimize AMD Instinct™ GPUs","url":"https://fireworks.ai/blog/fireworks-amd-instinct-gpu-partnership","excerpt":"Fireworks and AMD signed a multi-year deal to tune the Fireworks stack for AMD Instinct MI325X and MI355X GPUs, so you get faster, lower-cost inference."}},{"id":"clm_AB8DED23F015","predicate":"partnership","value":"LangChain","valid_from":"2023-10-02","valid_to":null,"published":"2023-10-02","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/langchain-playground-fireworks-models","tier":"primary","publisher":"Fireworks AI","published_at":"2023-10-02T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-31","quote":"We are excited to team up with LangChain to bring Fireworks.ai open-source models to the LangSmith Playground — completely free of cost.","url":"https://fireworks.ai/blog/langchain-playground-fireworks-models","excerpt":"Test prompts on Mistral 7B Instruct, Llama 2, and Code Llama in the LangSmith Playground via Fireworks, free when signed in. Follow the setup steps."}},{"id":"clm_572F3C5003C7","predicate":"released","value":"FireRouter with Opus","valid_from":"2026-09-28","valid_to":null,"published":"2026-09-28","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/introducing-firerouter-with-opus","tier":"primary","publisher":"Fireworks AI","published_at":"2026-09-28T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-29","quote":"Today, we’re introducing FireRouter with Opus , optimized for the Opus family and now available in both our CLI and, for the first time, as a standalone router model.","url":"https://fireworks.ai/blog/introducing-firerouter-with-opus","excerpt":"FireRouter with Opus is our first router model for coding. Send routine tasks to top open models and keep Claude Opus for the rest. Proven 57% lower cost per coding session at 98.1% of Opus accuracy."}},{"id":"clm_C7AAA22E0C22","predicate":"released","value":"Ember-1","valid_from":"2026-09-27","valid_to":null,"published":"2026-09-27","recorded":"2026-10-29","confidence":0.5,"status":"unverified","assertable":false,"why_not":"unverified","superseded_by":null,"sources":[{"url":"https://vercel.com/changelog/ember-1-from-fireworks-now-available-on-ai-gateway","tier":"secondary","publisher":"Vercel","published_at":"2026-09-27T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-29","quote":"Ember-1 from Fireworks is now available on AI Gateway . Ember-1 is a research preview reasoning model built on Kimi K3 for coding and agentic workflows.","url":"https://vercel.com/changelog/ember-1-from-fireworks-now-available-on-ai-gateway","excerpt":"Ember-1 from Fireworks now available on AI Gateway"}},{"id":"clm_0D788A888BB2","predicate":"released","value":"ARCv3","valid_from":"2026-09-23","valid_to":null,"published":"2026-09-23","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/arcv3-and-the-case-for-cross-region-rl","tier":"primary","publisher":"Fireworks AI","published_at":"2026-09-23T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-30","quote":"We’re releasing ARCv3, the newest version of the compressor we use to build those deltas.","url":"https://fireworks.ai/blog/arcv3-and-the-case-for-cross-region-rl","excerpt":"Fireworks’ ARCv3 delivers lossless BF16 weight updates nearly 50% smaller than ARCv2, making cross-region reinforcement learning more practical."}},{"id":"clm_94CB31DDB31C","predicate":"released","value":"Fireworks Nexus","valid_from":"2026-07-26","valid_to":null,"published":"2026-07-26","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireworks-nexus-open-model-routing","tier":"primary","publisher":"Fireworks AI","published_at":"2026-07-26T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-29","quote":"Available Today for Engineering Organizations: Fireworks Nexus Fireworks Nexus connects the AI tools your developers already use to a managed layer of open-weight models, with enterprise controls and intelligent routing.","url":"https://fireworks.ai/blog/fireworks-nexus-open-model-routing","excerpt":"Fireworks Nexus routes routine AI coding work to open models like GLM-5.2 and Kimi-K3, with team budgets and spend tracking. Keep the tools your engineers use."}},{"id":"clm_A8F83E9DD2BB","predicate":"released","value":"3D FireOptimizer","valid_from":"2025-06-14","valid_to":null,"published":"2025-06-14","recorded":"2026-10-30","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/3d-fireoptimizer-llm-serving","tier":"primary","publisher":"Fireworks AI","published_at":"2025-06-14T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-30","quote":"Now, we’re excited to announce a new toolkit in our FireOptimizer tuning stack: 3D FireOptimizer .","url":"https://fireworks.ai/blog/3d-fireoptimizer-llm-serving","excerpt":"3D FireOptimizer searches thousands of LLM serving configurations across hardware, quantization, and speculation to hit your quality, latency, and cost targets."}},{"id":"clm_BAD8FA1709DF","predicate":"released","value":"Experimentation Platform","valid_from":"2025-06-11","valid_to":null,"published":"2025-06-11","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/build-sdk-experimentation-platform","tier":"primary","publisher":"Fireworks AI","published_at":"2025-06-11T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-29","quote":"To help address these challenges, we’re excited to announce the GA of the Fireworks Experimentation Platform – designed to supercharge your experimentation velocity by reducing your iteration time from weeks to hours.","url":"https://fireworks.ai/blog/build-sdk-experimentation-platform","excerpt":"Run fine-tuning and eval experiments in Python with the Build SDK (Beta) and Experimentation Platform (GA). See code for deployments, LoRA, and setup."}},{"id":"clm_9A687B2C03E9","predicate":"released","value":"Fireworks f1","valid_from":"2024-11-15","valid_to":null,"published":"2024-11-15","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/f1-compound-ai-reasoning-model","tier":"primary","publisher":"Fireworks AI","published_at":"2024-11-15T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-29","quote":"Announcement NOV 15, 2024 Fireworks f1: a breakthrough in complex reasoning with compound AI","url":"https://fireworks.ai/blog/f1-compound-ai-reasoning-model","excerpt":"f1 interweaves multiple open models at the inference layer to solve complex math, coding, and logic problems. See worked examples and try f1 and f1-mini."}},{"id":"clm_79D359FFED61","predicate":"released","value":"FireOptimizer","valid_from":"2024-08-30","valid_to":null,"published":"2024-08-30","recorded":"2026-10-28","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/fireoptimizer-adaptive-speculative-decoding","tier":"primary","publisher":"Fireworks AI","published_at":"2024-08-30T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-29","quote":"Today, we’re excited to introduce FireOptimizer , our adaptation engine designed to customize and enhance AI model performance for your unique use cases and workload.","url":"https://fireworks.ai/blog/fireoptimizer-adaptive-speculative-decoding","excerpt":"FireOptimizer trains a draft model on your own traffic, so adaptive speculative execution cuts inference latency up to 3x compared with a generic draft model."}},{"id":"clm_DFC28A524A45","predicate":"released","value":"FireLLaVA","valid_from":"2024-01-18","valid_to":null,"published":"2024-01-18","recorded":"2026-10-29","confidence":0.7,"status":"valid","assertable":true,"superseded_by":null,"sources":[{"url":"https://fireworks.ai/blog/firellava-commercial-open-source-llava-model","tier":"primary","publisher":"Fireworks AI","published_at":"2024-01-18T00:00:00+00:00"}],"evidence":{"by":"page-read","on":"2026-10-29","quote":"We are excited to announce that we have open-sourced FireLLaVA under the Llama 2 Community License. It is the first LLaVA multi-modality model with a commercially permissive license.","url":"https://fireworks.ai/blog/firellava-commercial-open-source-llava-model","excerpt":"FireLLaVA is a 13B vision-language model under the Llama 2 Community License, trained on OSS-generated data. Download the weights or call it on Fireworks."}}],"legend":{"dates":"valid_from: when the claim took effect (a period figure: the day its period closed). published: the earliest day a source on the record published it (a filing's filing date, a post's date; null when no source carries one). recorded: the day AsOf first held it. A figure can take effect weeks before it is published.","why_not":{"unverified":"Unverified: one source that is not primary, or a primary source below confidence 0.7. Asserted once a primary source or a second independent domain confirms it."},"checked_by":{"page-read":"its source page states the value and the date"}}}