Model digest

Pick the right
model.

A weekly-refreshed digest of three picks per category — best, balanced, cheapest — with honest strengths, weaknesses, and a cost framing. Curated by a model that reads the live OpenRouter catalog every Sunday morning. Want the long tail? Browse every model →

Reasoning

Text models turn prompts into written answers: chat, drafts, code, summaries, and step-by-step reasoning.

Best

openai/gpt-6-astra-pro

Strengths
Top-tier analysis, software engineering, research, and long multi-step tasks.
Trade-offs
Very expensive; pro mode spends extra tokens and answers more slowly.

$10 / $50 per M tokens; pro uses more

Balanced

anthropic/claude-sonnet-5.5

Strengths
Reliable default for coding, bug fixes, and polished everyday writing.
Trade-offs
Trails flagships on the hardest research and long-horizon agent work.

$2 / $10 per M tokens

Cheapest

nvidia/nemotron-3.5-super-vl:free

Strengths
Free open model that handles chat, documents, and simple image questions.
Trade-offs
Rate limits and lower reliability; weaker reasoning than paid models.

$0 / $0 per M tokens

Image

Image models turn a text prompt into pictures, and many can also edit or restyle an image you already have.

Best

openai/gpt-5.4-image-2

Strengths
Top instruction following, photoreal detail, and text-in-image for hard prompts and multimodal edits.
Trade-offs
Most expensive option; slower and unnecessary for simple or high-volume jobs.

$8 / $15 per M tokens

Balanced

google/gemini-3.1-flash-image

Strengths
Near-pro image quality and editing at flash speed; strong everyday default.
Trade-offs
Trails the flagship on the hardest prompts, fine text, and deep visual reasoning.

$0.5 / $3 per M tokens

Cheapest

google/gemini-3.1-flash-lite-image

Strengths
Fast, low-cost drafts and bulk exploration that still look usable.
Trade-offs
Softer detail, weaker text rendering, and less reliable complex edits.

$0.25 / $1.5 per M tokens

Video

Models that create short video clips from a text prompt or a still image, sometimes with motion control and sound.

Best

google/veo-3.1

Strengths
Highest visual fidelity here, with 1080p output and native synchronized audio from text or image prompts.
Trade-offs
Costs more and runs slower than Fast or Lite Veo; clips stay short and it is not a general video editor.

$0.40 per second (varies by SKU)

Balanced

google/veo-3.1-fast

Strengths
Strong everyday quality with native audio, quicker and far cheaper than full Veo 3.1.
Trade-offs
A step below full Veo 3.1 on fine detail and motion; less ideal for final hero shots.

$0.12 per second (varies by SKU)

Cheapest

bytedance/seedance-1-5-pro

Strengths
Makes video and audio together in one pass at the lowest listed rate in this catalog.
Trade-offs
Older than Seedance 2.x; detail and motion trail flagships, and the tiny listed rate varies by SKU.

$0.0000024 per second (varies by SKU)

Last refreshed 8 hours ago. Auto-updates every Sunday at 03:00. Have a model we should add? Drop us a note.