MiniMax M3

minimax/minimax-m3

BackTry Model

MiniMax M3

minimax/minimax-m3

BackTry Model

MiniMax M3 is an open-weights frontier model built for coding, agent workflows, and multimodal understanding from step zero. MiniMax reports 59.0% on SWE-Bench Pro and 66.0% on Terminal Bench 2.1, with Sparse Attention designed to scale context to 1M. It starts with a 512K context cap for now, with 1M expected in the next week or so.

Added Jun 1, 2026

Context Window

512.0K

Max Output

80.0K

Input Price (Auto)

$0.31/1M

Output Price (Auto)

$1.26/1M

Cache Read (Auto)

$0.063/1M

Capabilities

Vision

Reasoning

Tool calling

Structured output

Benchmarks

Performance metrics and benchmarks

Artificial Analysis

LMArena

Sourced from Artificial Analysis.

No benchmark data is available yet for this model.

Providers

Choose explicit providers for this model. Auto routing remains available as the default option.

Loading provider options…