Back to Gap Map Model components / Inference code

Fireworks Inference

closed / Overall score: 4.0(strong)

Fireworks Inference runs open models on FireAttention, an in-house attention engine of custom CUDA and AMD-GPU kernels, combined with speculative decoding. It hosts many models and LoRA adapters concurrently, supports function calling, and offers both serverless per-token endpoints and dedicated GPU deployments. Its vendor, Fireworks AI, was founded by former Meta PyTorch engineers, including CEO Lin Qiao.

Openness

1 high confidence

1.0

engine: proprietary(FireAttention kernels)
source: closed
access: managed-cloud-API-only
license: Proprietary

A proprietary inference engine built on Fireworks' own FireAttention kernels and offered only as a managed cloud service. The fireworks.ai homepage carries no self-host, on-premises, VPC or BYOC option and no repository link anywhere on the page, and no LICENSE or license grant is published for the engine, so there is no source to run yourself and nothing to license.

https://fireworks.ai/ recorded 2026-09-23
the site markets Fireworks purely as a hosted API/pricing product (serverless and on-demand tiers, /pricing links); no self-host, on-prem, VPC, BYOC, or GitHub/source-repo link appears anywhere on the page

Adoption

4 medium confidence

4.0

Fireworks names a roster of high-volume production customers, including Cursor, Vercel, Notion, Sourcegraph, UiPath, Quora, Cresta and StackBlitz, each represented by a case study rather than a disclosed usage number. No customer discloses a request or user count, so it is that breadth of named customers, not a figure, that supports the rating.

https://fireworks.ai/ recorded 2026-09-29
The homepage carries customer stories naming Cursor, Vercel, Notion, Sourcegraph, UiPath, Quora, Cresta and StackBlitz; Notion says "we reduced latency from about 2 seconds to 350 milliseconds" and Quora says "we noticed a 3x speedup in response time".

Capability

4 medium confidence

4.0

Artificial Analysis measures Fireworks' fastest model among the 10 it lists, Nemotron 3.5 Lightning (BF16), at 355 output tokens per second. Fireworks has no entry on the gpt-oss-120b speed board, where Groq, SambaNova and Cerebras post 473 to 1,759 tokens per second on a different model, so the two sets of figures are not directly comparable.

https://artificialanalysis.ai/models/gpt-oss-120b/providers recorded 2026-09-29
The gpt-oss-120b board says "the top providers are Cerebras (1,759.4 t/s), SambaNova (709.4 t/s), and Groq (473.2 t/s)"; searched the body for Fireworks and found it only in the provider-logo list, not as a speed entry.
https://artificialanalysis.ai/providers/fireworks recorded 2026-09-29
The provider page lists 10 models and says "For output speed, the fastest models are Nemotron 3.5 Lightning (BF16) (355 t/s), DeepSeek V4.1 Flash (max) (347 t/s), and GLM-5.3 (max) (234 t/s)" with "a 97% difference between the fastest and slowest".
https://fireworks.ai/ recorded 2026-09-29
The homepage carries customer stories where Notion says "we reduced latency from about 2 seconds to 350 milliseconds" and Quora says "we noticed a 3x speedup in response time".

Verified 2026-09-23