Fireworks · Blogs

Blogs listings from Fireworks on TechiSeek. 148 public listings.

• DeepSeek V4 Pro: Validating Frontier Models for Production (Fireworks)
  DeepSeek V4 Pro is live on Fireworks — after full validation. See the reasoning trace bug we caught across day-0 endpoints before shipping…
• Qwen 3.7 Plus is now live on Fireworks (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• FireLLaVA: the first commercially permissive OSS LLaVA model (Fireworks)
  We have come to rely heavily on text as input for foundation models to generate responses.
• Launching Fireworks for Startups Program! (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Accelerating Code Completion with Fireworks Fast LLM Inference (Fireworks)
  At Fireworks.ai, we provide the world's fastest LLM inference platform which enables developers to run, fine-tune, deploy, and share large…
• Simplifying Code Infilling with Code Llama and Fireworks.ai (Fireworks)
  Llama 2 is one of the latest open-source foundation large language models (LLMs) from Meta AI.
• Kimi K3 is competitive with Fable; Kimi K3 + Fable is SoTA. (Fireworks)
  Compare Kimi K3 to Fable 5 on coding benchmarks, costs, and more. See results across 1,000+ agentic tasks, pricing breakdowns, and when to…
• Cursor Composer 2 + Fireworks (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Train past the frontier: Training API now generally available (Fireworks)
  Today, Fireworks is announcing the general availability of our Training API and Fireworks Lab, completing three paths designed to help team…
• Best Practices for Multi-Turn RL (Fireworks)
  Discover how Multi-Turn Reinforcement Learning (RL) is essential for building robust LLM agents capable of sequential tool use and complex,…
• DeepSeek V3 just got vision capabilities! (Fireworks)
  Learn how to use DeepSeek V3, the newest SOTA model released on Fireworks. Add vision capabilities to DeepSeek V3 with document inlining in…
• Vision Model Platform Updates: Enhanced Capabilities and New Features (Fireworks)
  Built on the same FireAttention serving stack that powers our LLMs, Fireworks delivers unmatched speed for real-time visual intelligence ap…
• Why do all LLMs need structured output modes? (Fireworks)
  Tl;dr: All Fireworks language models can now be invoked with either a (a) JSON schema (b) context-free grammar (similar to Llama.cpp’s feat…
• Frontier-lab training infrastructure, now as a service (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Introducing FireRouter with Opus (Fireworks)
  FireRouter with Opus is our first router model for coding. Send routine tasks to top open models and keep Claude Opus for the rest.
• GLM 5.2 Fast is live on Fireworks (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Introducing OpenAI gpt-oss (20b & 120b) (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• The Best 8 LLM API Providers in 2026 (Fireworks)
  We compared pricing, latency, and models across 8 LLM API providers. Find the best fit for production, fine-tuning, custom deployment, or p…
• Introducing Fireworks on Microsoft Foundry: Bringing Best-in-Class Open Model inference to Azure (Fireworks)
  This partnership integrates Fireworks’ high-performance inference and State-of-the-Art (SOTA) open models into the unified Microsoft Foundr…
• Accelerate your Vision Pipelines with the new NVIDIA Nemotron Nano 2 VL Model on Fireworks (Fireworks)
  This 12B vision language model offers leading accuracy for document intelligence and video understanding, transforming workflows from invoi…
• Can open models carry readable silent signals before they speak? Reproducing J-Lens Readouts on Kimi (Fireworks)
  Do AI models actually reason before they speak? We applied Anthropic’s J-Lens probe to open-weights models like Kimi K3 and Qwen 3.5-9B to…
• Introducing Supervised Fine-tuning V2 (Fireworks)
  We're excited to unveil Fireworks Supervised Fine Tuning Engine V2, a major upgrade that introduces more extensive model support, larger co…
• GPUs on-demand: Not serverless, not reserved, but some third thing (Fireworks)
  Note: Fireworks A100 and H100 prices have since been reduced to $2.90 and $5.80!
• Fireworks Nexus: Drop-in Open Frontier Intelligence for Teams with Budgets (Fireworks)
  Fireworks Nexus is a drop-in AI management and routing platform that helps engineering teams drastically cut software development costs by…
• Beyond Supervised Fine Tuning: How Reinforcement Learning Empowers AI with Minimal Labels (Fireworks)
  Deep Seek R1 and Deep Seek R1-Zero are all the rage right now. While Deep Seek R1 is likely a more suitable choice for production, Deep See…
• Your AI Performance Stack is Fireworks Models with Voyage AI embeddings (Fireworks)
  The full Voyage lineup now runs natively on Fireworks: the Voyage 4 family, voyage-multimodal-3.5, and rerank-2.5.
• The DeepSeek Model Lineup: V3.2, R1, and Distilled Variants Mapped to Production Workloads (Fireworks)
  DeepSeek models explained: V3.2 vs R1 vs distilled variants, when to use thinking mode, tool calling limits, token costs, and how to run th…
• Inference Providers vs. API Routers: Where Do Your Tokens Actually Come From? (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Frontier AI at a fraction of the cost: open-source worker agents with a closed-source advisor. (Fireworks)
  Boost your agentic AI performance while slashing costs with a hybrid "worker-advisor" architecture.
• Making the leap to specialized intelligence (Fireworks)
  The road to building specialized intelligence typically follows four stages from renting the closed frontier to training your own custom mo…
• Claude Code Pricing: Plans, API Costs, and How To Lower Your Bill (Fireworks)
  Compare Claude Code costs across Pro, Max, Team, Enterprise, and API usage, then learn how Fireworks can lower model costs for supported wo…
• Mixtral 8x7B on Fireworks: faster, cheaper, even before the official release (Fireworks)
  The newest Mistral AI MoE model, Mixtral 8x7B, is available on the Fireworks platform in both base and instruction-tuned variants.
• Fireworks Raises $52M Series B to Lead Industry Shift to Compound AI Systems (Fireworks)
  Fireworks raises $52M led by Sequoia, boosting its valuation to $552M. The funds will speed up the development of compound AI systems, team…
• How Factory Grew Open Model Usage 2-3x in Six Months on Fireworks (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Build for Scale with Fireworks Virtual Cloud (GA) (Fireworks)
  We’re excited to announce the GA of the Fireworks Virtual Cloud, a platform that abstracts away the complexity of managing GPU deployments,…
• Building a RAG with Astro, FastAPI, SurrealDB and Llama 3.1 (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Speed, Python: Pick Two. How CUDA Graphs Enable Fast Python Code for Deep Learning (Fireworks)
  This is the second in a series of technical blog posts about the techniques we use for optimization of the high-performance Fireworks Gen A…
• Fireworks Now Supports Amazon SageMaker (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Build Your Own Flight Recommendation System using FastAPI, SerpAPI, and Firefunction (Fireworks)
  A recommendation system for flights using Fireworks Firefunction-v2, SerpApi, FastAPI, and Next.js.
• DeepSeek-V4.1-Flash on Fireworks: Astra-level DeepSWE at 1/15th the cost (Fireworks)
  DeepSeek-V4.1-Flash hits GPT-6 Astra-level coding accuracy on DeepSWE at $0.43 per task — 15x cheaper.
• Enabling Function Calling in DeepSeek v3: Bridging the Gap Between Text and Action (Fireworks)
  Large language models (LLMs) generate impressive text but cannot access or update real-world information.
• LLM on the edge: Model picking with Fireworks Eval Protocol + Ollama (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Open, frontier, and yours: LangChain Deep Agents on NVIDIA Nemotron 3 Ultra, running on Fireworks (Fireworks)
  LangChain has tuned its Deep Agents harness for NVIDIA Nemotron 3 Ultra, achieving benchmark-leading agent performance among open models at…
• Fireworks Real-World Benchmarks: Find the Best OSS Model for the Job (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Fireworks.ai Achieves SOC 2 Type II and HIPAA Compliance (Fireworks)
  We are pleased to report that the Fireworks.ai inference platform is both SOC 2 Type II and HIPAA compliant.
• LLM Inference Performance Benchmarking (Part 1) (Fireworks)
  Optimizing Large Language Model (LLM) machine performance in inference is a complex space and no solution is one-size-fits-all.
• FireAttention (Fireworks)
  Mixtral has recently made waves in the AI community as the first OSS model trained on trillions of tokens to support 'mixture of experts' (…
• Fireworks AI (Fireworks)
  Fireworks has acquired Hathora, and we're thrilled to bring their team and technology into the Fireworks family.
• Your AI Benchmark is Lying to You. Here's How We Caught It (Fireworks)
  Use state-of-the-art, open-source LLMs and image models at blazing fast speed, or fine-tune and deploy your own at no additional cost with…
• Make Kimi K3 Yours: LoRA Training on Fireworks (Fireworks)
  Fireworks now supports private preview access for Serverless LoRA Training on Kimi K3, the world’s first open-weight 2.8-trillion parameter…