ku
Urugero: ni ya: Igikorwa? i Kuri Kabiri na i in Ubwiza, Umuvuduko, na Imisusire., Oya.
Gukoresha
1. Kabiri Imisusire
Choose from 27 models across 3 tiers: free self-hosted, standard, and premium. Try Qwen vs GPT, Claude vs Gemini, or any combination.
2. a
Type your question or task. It goes to both models simultaneously. Each gets its own conversation history so you can have multi-turn comparisons.
3. Compare results
Iburyo: ku Iburyo: Na:. Urugero: ya: Koresha. mu buryo bwikora:.
Gutandukanya
GPT-4.1 vs Claude Sonnet 4.6
ya: Na
ya: Na
Qwen 3 30B vs Mistral 7B
Free self-hosted showdown
Free self-hosted showdown
GPT-5 vs Gemini 2.5 Pro
Ipaji Kuri Ipaji
Ipaji Kuri Ipaji
DeepSeek R1 vs o4 Mini
Urugero
Urugero
GPT-4.1 Mini vs Llama 4 Maverick
Agaciro hagati
Agaciro hagati
Grok 4 vs Claude Opus 4.6
Ibara ry' Ipaji:
Ibara ry' Ipaji:
Bihari
| Itariki | Amadosiye | Igihe cyarenze: | Byiza |
|---|---|---|---|
| Ibiro | Qwen 3 30, Qwen 3 8, Mistral 7, DeepSeek R1 7, Qwen | <B B | Inyandikonyakwirema Inyandikoporogaramu |
| Bisanzwe | GPT-4.1 Mini, GPT-5 Mini, Gemini 2.5 Flash, Llama 4, DeepSeek V3 | ~ | Ibikorwa, Ibigize, |
| Icyiciro gikomeye | GPT-4.1, GPT-5, Claude Sonnet/Opus 4.6, Gemini 2.5 Pro, Grok 4 | ~1. - | Igerageza, |
Ibibazo bizwa kenshi
Na: Kabiri ku. Rimwe & Ubutumwa na - Guhuza Ubwiza, Umuvuduko, na Imisusire.
Compare any two of our 27 models: 6 free self-hosted models (Qwen, Mistral, DeepSeek), 10 standard models (GPT-4.1 Mini, Llama 4, Gemini Flash), and 11 premium models (GPT-5, Claude 4.6, Grok 4). Mix and match freely.
Yes! Each response costs tokens based on the model used. Self-hosted models cost ~100 tokens/msg, premium models cost more. You get 6,000 tokens/day free, or sign up for 30,000/day
Yes - each model response is billed independently. Comparing two models costs roughly double a single chat message. The token cost is shown below each response.
! Urugero: Urutonde. - Hejuru Ubutumwa na Byombi Urugero: Kwibuka Ibanjirije. Gugereranya, Igikubo - Intera Ibikorwa:.
. in Mucukumbuzi na Gihinguranya Ipaji:. Hindura hagati ikoresha i Umukoro Amasimbuka hejuru i. Kuri 20.
Self-hosted models (Qwen, Mistral) run on our own GPUs - cheapest at ~100 tokens/msg and your data stays on our servers. Premium models (GPT-5, Claude, Gemini) are from external providers and cost more but offer stronger reasoning. Compare them to decide what is worth the extra cost for your needs.
Model Compare lets you choose which models to compare and have full conversations. The AI Arena (/arena/) is blind - you don't know which models you're comparing until after you vote. Arena builds a community ELO leaderboard. Use Compare for personal evaluation, Arena for community rankings.