GPT-4.1 Mini against 49 models

GPT-4.1 Mini costs $1.44 per 1,000 chat replies at list price. Here is every model teams spend the most on or move traffic to, sorted from cheapest, with a link to each side-by-side page.

GPT-4.1 Mini compared with 49 models
ModelPer 1,000 chat repliesTo scaleAgainst GPT-4.1 MiniConfidence signal
gpt-oss-20b$0.07295% cheaperyes
Qwen3.7 Flash$0.11292% cheaperyes
Nova Micro 1.0$0.12691% cheaperno
Command R7B (12-2024)$0.13591% cheaperno
DeepSeek V4 Flash 0731$0.1788% cheaperyes
Phi 4$0.19686% cheaperno
Nova Lite 1.0$0.21685% cheaperno
Nemotron 3.5 Lightning$0.2483% cheaperyes
DeepSeek V4.1 Flash$0.2980% cheaperyes
Gemma 4 31B$0.31678% cheaperyes
Llama 3.3 70B Instruct$0.32877% cheaperyes
Ministral 3 8B 2512$0.3675% cheaperno
MiMo-V2.6-Flash$0.39273% cheaperno
GPT-6 Luna$0.4072% cheaperno
Qwen3.8 Flash$0.48866% cheaperyes
GLM 5.3 Flash$0.5065% cheaperyes
gpt-oss-120b$0.5463% cheaperyes
Mistral Small 4$0.5463% cheaperno
Llama 4 Maverick$0.63656% cheaperyes
GPT-5.6 Luna$0.8839% cheaperno
GPT-5.4 Nano$0.9038% cheaperno
MiniMax M3$1.0825% cheaperyes
Mistral Large 3 2512$1.601.1× the priceno
Gemini 3.5 Flash Lite$1.601.1× the priceno
Qwen3.8 27B$2.041.4× the priceyes
Nemotron 3 Ultra$2.161.5× the priceno
Kimi K2.6$2.661.9× the priceyes
Hy4 preview$2.671.9× the priceno
DeepSeek V4 Pro 0423$2.671.9× the priceyes
Gemini 3.8 Flash$3.002.1× the priceno
GPT-5.4 Mini$3.302.3× the priceno
Claude Haiku 4.5$4.002.8× the priceno
GLM 5.3$4.563.2× the priceyes
Grok 4.7$5.123.6× the priceyes
Qwen3.8 Max (0902)$6.404.4× the priceyes
Claude Sonnet 5$8.005.6× the priceno
GPT-6 Sol$8.005.6× the priceno
GPT-5.6 Sol$8.005.6× the priceno
GPT-5.6 Terra$8.806.1× the priceno
Gemini 3.1 Pro Preview$8.806.1× the priceno
Command A$9.006.3× the priceno
Nova Premier 1.0$10.006.9× the priceno
GPT-5.4$11.007.6× the priceno
Kimi K3$12.008.3× the priceyes
Claude Opus 5.5$16.0011× the priceno
Claude Opus 5$20.0014× the priceno
GPT-5.5$22.0015× the priceno
Claude Fable 5.1$40.0028× the priceno
GPT-6 Astra$40.0028× the priceno

Find the cheapest model that answers your GPT-4.1 Mini traffic as well

A pilot tests the candidates on your own requests.

Request a pilot