Quick
Fast and cheap, for drafts and short posts
Runs on GPT-6 Luna
AI model benchmarks
Every model in Nuvora,
judged on public scores.
The AI models you can pick in Nuvora, for text, images and video, measured on public benchmarks. Choose on facts, not habit.
Balanced
Good quality at a fair cost, for most work
Runs on Claude Sonnet 5.5
Best
The strongest, for important work and hard questions
Runs on Claude Opus 5.5
Scores: Text
The text models you can pick in Nuvora. Click a column to sort by it.
No model from this provider.
- Price level, from $ to $$$$, ranks the text models against one another and is not a price.
- vReported by the model’s maker
- No public figure; we never estimate one
What the benchmarks measure: Text
A high score on one test does not make a model better at everything. For marketing copy, look first at the overall index and at the arena, where people vote for the answers they prefer.
- Intelligence Index
-
One overall score that blends ten hard tests of reasoning, knowledge, coding and agent work. The quickest way to compare general ability.
Artificial Analysis (opens in a new tab) · Version 4.3, ten tests, each model at its strongest reasoning setting
- LMArena Text
-
A rating built from millions of blind votes in which people pick the better of two answers. It shows whose answers people prefer.
LMArena (opens in a new tab) · Overall text ranking, with style control
- Output speed
-
How many tokens, roughly words, the model writes per second once it starts answering. Higher feels faster.
Artificial Analysis (opens in a new tab) · Median speed measured by Artificial Analysis
- Humanity's Last Exam
-
About 2,000 very hard expert questions across many fields. Even the best models score well under 100%.
Artificial Analysis (opens in a new tab) · Without tools
- GPQA Diamond
-
Graduate-level science questions that are hard to answer even with a web search. A test of expert reasoning.
Artificial Analysis (opens in a new tab) · No longer run on the newest models, which show a dash
- tau2-bench Telecom
-
A customer support conversation in which the model must use tools and follow the rules correctly.
Artificial Analysis (opens in a new tab) · No longer run on the newest models, which show a dash
Scores: Image
The image models you can pick in Nuvora. Click a column to sort by it.
No model from this provider.
- Price level, from $ to $$$$ per image, ranks the image models against one another and is not a price.
- vReported by the model’s maker
- No public figure; we never estimate one
What the benchmarks measure: Image
A high score on one test does not make a model better at everything.
- AA Text to Image
-
How often people prefer this model's pictures over other models' in blind side by side votes on the same text prompt.
Artificial Analysis (opens in a new tab) · Overall board, all categories
- Arena Text to Image
-
Score from public blind votes comparing two models' pictures for the same prompt.
LMArena (opens in a new tab) · Overall board
- AA Image Editing
-
How often people prefer this model's edits of a supplied picture in blind side by side votes.
Artificial Analysis (opens in a new tab) · Overall board, all categories
- Arena Image Edit
-
Score from public blind votes comparing two models' edits of the same picture.
LMArena (opens in a new tab) · Overall board
- Generation time
-
Typical number of seconds the model takes to deliver one picture.
Artificial Analysis (opens in a new tab) · Median on the maker's own API, moves from day to day
Scores: Video
The video models you can pick in Nuvora. Click a column to sort by it.
No model from this provider.
- Price level, from $ to $$$$ per clip, ranks the video models against one another and is not a price.
- vReported by the model’s maker
- No public figure; we never estimate one
What the benchmarks measure: Video
A high score on one test does not make a model better at everything.
- AA Text to Video
-
How often people prefer this model's silent clips over other models' in blind side by side votes on the same text prompt.
Artificial Analysis (opens in a new tab) · Without audio
- Arena Text to Video
-
Score from public blind votes comparing two models' clips for the same text prompt.
LMArena (opens in a new tab) · Overall board
- AA Text to Video, with audio
-
How often people prefer this model's clips with sound over other models' in blind side by side votes on the same text prompt.
Artificial Analysis (opens in a new tab) · With audio
- AA Image to Video
-
How often people prefer this model's silent animation of a supplied picture in blind side by side votes.
Artificial Analysis (opens in a new tab) · Without audio
- Arena Image to Video
-
Score from public blind votes comparing two models' animation of the same picture.
LMArena (opens in a new tab) · Overall board
Start here
Ready to make LinkedIn
pay back?
Tell us about your company. We'll show you where the opportunity sits, and what it would take to capture it.
Thirty minutes. No pitch. Straight answers.