WWhichAITry via Vercel AI Gateway
Best for/cheap-bulk

Cheapest AI models for bulk workloads

When you're running millions of requests, value per million tokens swamps every other axis. Here are the strongest cheap models in 2026.

The podium

Top 3 picks

πŸ₯‡Anthropic
Claude Haiku 4.5

Fast, cheap, surprisingly smart.

value
10.0
speed
10.0
200K context$1.00 / $5.00
πŸ₯ˆGoogle
Gemini 3 Flash

Cheap, fast, 2M-token context.

value
10.0
speed
10.0
2M context$0.30 / $1.20
πŸ₯‰Google
Gemini Nano 2

On-device first. Free inference, private by default.

value
10.0
speed
10.0
32K contextFree / Free
Full ranking

Runners-up

#ModelContextPrice /M
4GPT-5 Mini OpenAI200K$0.40 / $1.60View β†’
5Mistral Small 3 Mistral128K$0.20 / $0.60View β†’
6GPT-OSS 20B OpenAI128K$0.15 / $0.60View β†’
7Qwen 3 Coder Alibaba128K$0.50 / $1.50View β†’
8DeepSeek V3 DeepSeek128K$0.27 / $1.10View β†’
Keep learning

Related guides