Low cost

Mistral AI May 31, 2026

Mistral Small 4

Efficient instruct

Reasoning 82.0 Coding 86.0 256k tokens

DeepSeek May 31, 2026

DeepSeek-V4-Flash

Very low-cost long-context API workflows where DeepSeek compatibility and economics are attractive.

Reasoning 84.0 Coding 85.0 1M tokens

Mistral May 20, 2026

Devstral 2

Code agents

Reasoning 82.0 Coding 92.0 256K tokens

DeepSeek May 20, 2026

DeepSeek V4 Pro

Cost-sensitive advanced reasoning and coding where context size still matters.

Reasoning 89.0 Coding 89.0 1M tokens

OpenAI May 20, 2026

GPT-5.4 mini

High-volume agent backends

Reasoning 86.0 Coding 90.0 400K tokens