Product / Model Selection
The Vera
Model Garden
The best models for the job, curated. Strong defaults, full transparency, zero decision fatigue.
Best practice comes free on Vera
Modes over models.
The right model for the job. The Job To Be Done.
You wouldn't take a Ferrari camping. You wouldn't drive a Humvee to the grocery store. Always using the biggest, most expensive model isn't the move — it's slow, it's expensive, it's unreliable.
Even the foundation model companies know this — their big models are Mixture of Experts under the hood anyway.
Each mode defaults to the best curated model for the task. Planning needs deep reasoning. Building needs speed and reliability. Quick iteration needs instant responses. Different tasks, different defaults — and you can change any of them.
Defaults change as models get better. Swap any of them in settings, or bring your own keys.
Modes
All defaults are instantly swappable on Vera Cloud. Bring your own keys or use ours.
Vera · Interview mode · Kimi K2.5
Before I start asking questions — which directories or files should I look at for context?
You
The API is in packages/api and the frontend is in packages/app. We just migrated from a monolith.
Vera · Interview mode · Kimi K2.5
Got it — I’ll only look there. Question 1: What’s the main goal? A) New feature from scratch B) Redesigning something existing C) Fixing/improving what’s there D) Something else
Today's Model Menu · April 15, 2026
Every model. Full transparency.
Swap any default instantly on Vera Cloud. Here's everything we have access to.
GPT 5.3 Codex
OpenAI
OpenAI’s answer to Opus. Built specifically for code.
Claude Opus 4.6
[April 2026] Being nerfed — quality declining since last update
Anthropic
The best coder alive. Slow and expensive, but worth it.
Gemini 3 Pro
Google’s heavyweight. Million-token context window.
MiniMax M2.7
MiniMax
Chinese dark horse. Punches way above its price.
Claude Sonnet 4.6
[April 2026] Super slow — constant queuing
Anthropic
90% of Opus brains at half the price.
GPT 5.1 Codex Mini
OpenAI
Budget Codex. Handles most coding tasks just fine.
Gemini 3 Flash
Fast and capable. Best bang-for-buck from Google.
Z.AI GLM 4.7
Cerebras
1000 tokens per second. Absurdly fast.
~1,000 tok/s
GPT OSS 120B
Cerebras
130x cheaper than Opus. 3,000 tok/s.
~3,000 tok/s
Qwen 3 235B
Cerebras
Biggest brain on Cerebras hardware.
~1,400 tok/s
Llama 4 Maverick
Groq
Meta’s open model. Sees images.
~500 tok/s
Kimi K2.5
Kimi
Most human-sounding model. Great at visual coding.
Flux Pro
fal.ai
Best photorealistic output right now. Not close.
Nano Banana Pro
Generates and edits images. Text-to-image built in.
Kling v3
fal.ai
Most consistent video gen we’ve tested.
Veo 3
Google’s video model. Strong on cinematic style.