Puter Developer Blog
News and updates from the Puter Engineering Team
GPT-6 Astra Review: A Taste of AGI
OpenAI just released GPT-6 Astra, which it calls the world's most intelligent and aligned model, with saturating scores on ARC-AGI-3, FrontierMath, and ExploitBench. We look at what is actually new and test one of its claims ourselves.
Gemini 3.8 Flash Review: A Diligent Model That Works Harder
Gemini 3.8 Flash matches Gemini 3.7 Flash on price but posts better benchmarks, which Google credits to the model being more diligent. We looked at the performance numbers and the diligence claim, and tested it ourselves.
Claude Fable 5.1 Review: Better Results with Less Effort
Claude Fable 5.1 is out. We put together a cost estimate of what its lower-effort, lower-cost design means for coding work.
Apodex 1.1 Review: Research With a Built-In Verification Step
Apodex says its Statement Review step independently checks key claims before delivery. We fed Apodex 1.1 and Apodex 1.1 Mini a prompt with a deliberately wrong gross margin figure and watched what they did with it.
GLM-5.3 Review: Stronger Coding With Fewer Tokens
Z.ai says GLM-5.3 beats GLM-5.2 with fewer output tokens. We ran 8 coding tasks against both models via Puter.js at max reasoning effort and measured the actual token spend.
Gemini 3.7 Flash Review: Google's Workhorse Model
Google says Gemini 3.7 Flash follows instructions with greater fidelity and needs less manual oversight than 3.6 Flash. We ran 12 single-turn tasks against both models through Puter.js and checked the outputs against binary pass criteria.
Grok 4.6 Review: xAI's Stronger Visual and Interactive Model
xAI says Grok 4.6 produces stronger first passes on visual and interactive projects than Grok 4.5. We built the same app with both and graded what came out.
Introducing BytePlus Models “Seedance, Seedream, and Seed” on Puter.js in Collaboration With the BytePlus Team
BytePlus, ByteDance's model platform, is now integrated directly with Puter, bringing the Seed, Seedream, and Seedance families to Puter.js, with new models added as they launch.
Announcing Puter AI Builder
We're launching AI Builder, a tool that lets you build apps using plain language, powered by Puter.js. Apps are secure by default, cost nothing for developers to run, and publish to their own link in one click.
Kimi K3 Review: Benchmarks, Pricing, and a Visual Coding Test
We reviewed Kimi K3, Moonshot AI's 2.8T open-weight model, a Fable 5-level model with Sonnet 5 pricing. Reported benchmarks vs independent numbers, the $3/$15 pricing math, and a hands-on visual coding test.
The Best Backend Platform for Vibe Coding
A deep dive on choosing the best backend platform for vibe coding. We compare Puter.js, Firebase, and Supabase on MCP support, safe defaults, and more, and share the rules files that keep AI code correct.
Kwaipilot KAT-Coder V2.5 Models Are Now Available in Puter.js
Access KAT-Coder-Pro V2.5 and KAT-Coder-Air V2.5, Kwaipilot's agentic coding models, through Puter.js for free. No API keys required.
Nano Banana 2 Lite Review: Testing Google's "Fastest, Most Cost-Efficient" Claim
We tested Google's speed and cost claims for Nano Banana 2 Lite, timing it against Nano Banana 2, Nano Banana Pro, and the original Nano Banana.
Best Netlify Alternatives (2026)
Discover the best Netlify alternatives, including Puter, Cloudflare Pages, GitHub Pages, Vercel, Firebase Hosting, and Tiiny.host. Compare features, pricing, and find the best fit for your project.
Best AI Gateway in 2026
Compare the best AI gateways of 2026: Puter.js, TrueFoundry, OpenRouter, LiteLLM, Cloudflare AI Gateway, Portkey, and Kong. Side-by-side features, pricing, and use cases.
Claude Sonnet 5 Review: Testing Anthropic's "Most Agentic" Claim
We tested Anthropic's agentic claims for Claude Sonnet 5, covering what agentic means in practice, our hands-on results, and how to add it to your app for free with Puter.js.
Testing Ideogram 4.0's Text Rendering and Native 2K Output
Ideogram 4.0 is now in Puter.js. We tested its text rendering and native 2K output with a prompt built to be checked string by string.
Poolside Laguna XS 2.1 Is Available in Puter.js
Laguna XS 2.1, Poolside's updated compact agentic coding model with improved multilingual and terminal-task performance, is available through Puter.js—no API keys required.
Grok Imagine Image Review: Testing xAI's 15-20x Faster Claim
Grok Imagine Image and its Quality tier are now in Puter.js. See how their speed and price compare to other flagship image models.
Sakana Fugu Ultra Is Now Available in Puter.js
Access Fugu Ultra, Sakana AI's multi-agent orchestration model that routes tasks across a pool of frontier LLMs and scores 93.2 on LiveCodeBench and 95.5 on GPQA-Diamond, through Puter.js—no API keys required.