Rankings
What people prefer, by the kind of work asked, beside what each model costs and how fast it answers.
Reasoning
No votes in this category yet.
No votes here yet
No one has voted on a reasoning prompt yet. <a href="/arena">Ask one</a>.
Products and apps
Which models OpenVibe’s products and developers’ apps run (apps as one row, never named), from AI’s own aggregate usage. No prompt and no person.
| Requester | Models | Requests |
|---|---|---|
 OpenVibe.Live | gpt-5-nano-2025-08-07, gpt-5-mini-2025-08-07, ggml-parakeet-tdt-0.6b-v3-q8_0.bin | 27,513 |
 OpenVibe.Trade | gpt-5-nano-2025-08-07 | 2,665 |
| OpenVibe.Chat | gpt-5-nano-2025-08-07, qwen2.5-1.5b-instruct, deepseek-flash | 1,236 |
| OpenVibe.Meme | gpt-5-nano-2025-08-07 | 344 |
| OpenVibe.Search | bge-small-en-v1.5, megasource, text-embedding-3-small | 261 |
| OpenVibe.News | gpt-5-nano-2025-08-07 | 256 |
| OpenVibe.Blog | gpt-5-nano-2025-08-07 | 221 |
| OpenVibe.Food | gpt-5-nano-2025-08-07 | 76 |
 OpenVibe.Games | llama-3.2-1b-instruct, qwen2.5-1.5b-instruct | 37 |
| OpenVibe.Wiki | gpt-5-mini-2025-08-07, megasource | 25 |
| OpenVibe.Bench | gpt-5-nano-2025-08-07 | 20 |
 OpenVibe.Video | qwen2.5-1.5b-instruct, gpt-5-nano-2025-08-07, llama-3.2-1b-instruct | 13 |
How to read it
- Rating is Bench Elo from blind Arena votes. 1000 is where a model starts; a model with fewer than 10 games is provisional.
- Price is the cheapest live offer’s price per million tokens, in and out, from OpenVibe.AI’s catalogue.
- Speed and uptime are measured from real requests over the last seven days; an offer with no traffic shows a dash.
- Value is rating points above 1000 per dollar of output. It rewards a model that wins and costs little; it never turns a dear model into a bargain. A model whose price is not published shows a dash.
- Categories come from the prompt’s classification in the Arena: code, writing, chat, reasoning, translation and analysis.