AI Tools
Fireworks AI
Fireworks AI: preços, capacidades, atualizações e alternativas, lidos nas páginas do próprio desenvolvedor.

Em inglês
Key facts
- Price
- Paid plans: see pricing Read on fireworks.ai
- Free plan
- None listed on its pages
- Platforms
- Web · CLI · API
- API
- Yes Read on fireworks.ai
- Open source
- Not presented as open source
- Certifications
- SOC 2 Type II · ISO 27001 · ISO 42001 · GDPR · HIPAA Read on trust.fireworks.ai
- Last update
- 4 Nov 2026 Read on docs.fireworks.ai
- Prices checked
- 8 Nov 2026 Read on fireworks.ai
Fireworks provides model inference through an API, with serverless and dedicated GPU deployment options. Teams can route requests across models, combine models, and train open models on their own data.
Developers and ML teams use it to build and operate AI applications, agents, and customized models. Usage is billed by tokens or GPU time; reserved capacity, training, enterprise contracts, and Nexus plans may cost extra. The platform is not itself an open-source software product.
- Category
- Developer Tools
- Read from the maker’s pages
- 2026-11-18
Company
Features
- Serve open and custom models through an API
- Choose serverless or dedicated GPU deployments
- Route requests across models
- Combine models for applications and team workflows
- Fine-tune open models on your data
- Train with token-based or GPU-hour options
- Run Fireworks-managed inference in your own cloud
Use cases
- Build an application that calls hosted models through an API
- Deploy a model for steady production traffic with reserved throughput
- Fine-tune an open model on company data
- Route routine agent tasks to lower-cost models
- Run inference in your cloud to meet data-residency needs
Pricing
Paid plans: see pricing
- Prices checked
- 2026-11-08
Serverless
From $0.05 / 1M input tokens
- From $0.05 / 1M input tokens
- Pay per token; no commitment
- Standard, Priority, Fast, Batch and Spot modes
- Cached input billed at half price or less
Reserved Throughput
The maker’s pricing page names this plan. The page we read prints no price for it. See the maker’s pricing page
Serverless Training API
The maker’s pricing page names this plan. The page we read prints no price for it. See the maker’s pricing page
We build with you
The maker’s pricing page names this plan. The page we read prints no price for it. See the maker’s pricing page
Fireworks Sovereign
price on request
- Fireworks-managed inference in your cloud account or data center
- Inference and response handling stay inside your boundary
- Optional overflow to Fireworks-managed capacity
Managed Training
From $0.50 / 1M training tokens
- From $0.50 / 1M training tokens
- No-code, just bring a dataset
- Launch SFT, DPO or ORPO from the UI, CLI or API
- Fireworks handles scheduling, training and checkpoints
- LoRA, or full-parameter on select models
Dedicated Training API
From $8.00 / GPU-hour
- From $8.00 / GPU-hour
- You run the loop, at full scale
- Provisioned trainers and deployments
- Full-parameter training, the largest MoE models, longer context and higher LoRA ranks
Pro
$8/mo
- $8 / month
- The right model for each task
- Open models included
- Usage against your plan cap
- Works with your AI tools
Team
$10/user/mo
- $10 per seat / month
- Everything in Pro
- Up to 200 seats
- Members, seats and billing
- Higher usage per seat
Enterprise
price on request
- SSO, SCIM and audit logs
- Groups and Model Policy
- Account, user and group limits
- Custom terms and support
Capabilities
- Command line — “API, SDK, CLI, console” source
- Choice of models — “Intelligently orchestrate models for work and applications.” source
- Self-hosted — “Yes. Fireworks Sovereign brings our serving stack to enterprise infrastructure in your own cloud account or data center.” source
- No code retained — “Retention depends on the features you use, with controls to manage stored data and enforce zero data retention for supported workloads.” source
- API — “API, SDK, CLI, console” source
- Official SDKs — “API, SDK, CLI, console” source
- Runs models for you — “Serve open and custom models” source
- Traces and evaluates — “For core workloads that encode your intellectual property, evaluate models against your own benchmarks and train specialized models to outperform general-purpose alternatives on the tasks that matter to your product.” source
Latest updates
- Cookbook migration and first PyPI release
The public cookbook now lives at fw-ai-external/cookbook; fireworks-training-cookbook 2026.11.5.1 is available on PyPI.
- New Serverless mode: Ultra-fast (Preview)
Ultra-fast is a new high-speed Serverless mode, now in Preview, for highly latency-sensitive applications.
- New Serverless mode: Ultra-fast (preview)
Ultra-fast is available in Preview for DeepSeek V4.1 Flash 1017.
- Serverless Training deprecation: DeepSeek V4 Flash (0731)
DeepSeek V4 Flash (0731) is removed from the shared Serverless Training pool effective November 2, 2026.
- New model on Serverless: Rune 26B-A4B
Rune 26B-A4B is now available on Serverless.
Where it works
Web · CLI · API
Security
Alternatives to Fireworks AI
- Together AI
Together AI provides APIs and infrastructure for running, fine-tuning, and training open AI models.
- Baseten
Baseten provides infrastructure to deploy and run AI models, train models, and build agent workflows using managed inference, APIs, training tools, and sandboxes.
- Cerebras Inference
Cerebras Inference serves AI models through a cloud API and web playground, with pay-per-token access for developers and reserved options for enterprises.
- Core42
Core42 provides sovereign cloud and AI infrastructure for training, deploying and running AI models, including inference through its Compass API.
- Replicate
Replicate lets developers run, fine-tune, and deploy AI models through a web interface or API.
- Lambda GPU Cloud
Lambda GPU Cloud rents NVIDIA GPU instances and clusters for training, fine-tuning, and serving AI models.
Questions about Fireworks AI
Is Fireworks AI free?
No free plan is listed on the pages read for Fireworks AI. Read on the maker’s pages on 18 Nov 2026.
How much does Fireworks AI cost?
The plans on the maker’s pricing page: Serverless (From $0.05 / 1M input tokens), Fireworks Sovereign (price on request), Managed Training (From $0.50 / 1M training tokens), Dedicated Training API (From $8.00 / GPU-hour), Pro ($8/mo), Team ($10/user/mo) and Enterprise (price on request). The page we read prints no price for Reserved Throughput, Serverless Training API and We build with you. Prices checked on 8 Nov 2026.
Does Fireworks AI have an API?
Yes. The maker’s pages state that Fireworks AI has an API. Read on the maker’s pages on 18 Nov 2026.
Which platforms does Fireworks AI run on?
Fireworks AI runs on Web, CLI and API, as its maker’s pages state. Read on the maker’s pages on 18 Nov 2026.
Is Fireworks AI open source?
The maker’s pages do not present Fireworks AI as open source. Read on the maker’s pages on 18 Nov 2026.
Is Fireworks AI still available?
Yes, as of 18 Nov 2026: its maker’s pages were read that day and carried no notice that Fireworks AI has ended.
What changed recently in Fireworks AI?
The latest entry in its maker’s changelog: 4 Nov 2026, “Cookbook migration and first PyPI release”. Before it: 3 Nov 2026, “New Serverless mode: Ultra-fast (Preview)”. Changelog checked on 15 Nov 2026.
Sources
- https://fireworks.ai/pricing — Pricing page read on 2026-11-18; lists usage, training, deployment, and Nexus plan prices.
- https://fireworks.ai/product — Product overview describes inference, model routing, training, API, CLI, and cloud deployment.
- https://fireworks.ai/blog?blogTypes=Announcement — Announcement listing identifies the November 3, 2026 Forge product announcement as the latest listed item.
Not confirmed
- Latest version
- The official product and pricing pages read do not identify a single current Fireworks platform version or model name.