Private AI
MiniMax M3 is the non-thinking route for MiniMax's open-weights frontier model, built for coding, agent workflows, tool use, and multimodal understanding from step zero. It keeps native thinking disabled for faster direct answers. 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 on NanoGPT for now.
Added Jun 1, 2026
Model weightsContext Window
512.0K
Max Output
80.0K
Avg output tokens (7d)
1.8K tokens
Input Price (Auto)
$0.31/1M
Output Price (Auto)
$1.26/1M
Cache Read (Auto)
$0.063/1M
Capabilities
Performance metrics and benchmarks
Sourced from Artificial Analysis.
Intelligence Index
44.4
Choose explicit providers for this model. Auto routing remains available as the default option.
Loading provider options…
Coding Index
58.6
Agentic Index
35.4
GPQA Diamond
Graduate-level scientific reasoning
92.9%
Better than 98% of models compared
HLE
Humanity's Last Exam
37.1%
Better than 94% of models compared
IFBench
Instruction-following benchmark
82.9%
Better than 99% of models compared
T²-Bench Telecom
Conversational AI agents in dual-control scenarios
88.9%
Better than 83% of models compared
AA-LCR
Long context reasoning evaluation
74.0%
Better than 98% of models compared
GDPval-AA
Economically valuable tasks
44.5%
CritPt
Research-level physics reasoning
3.7%
SciCode
Python programming for scientific computing
45.4%
Better than 84% of models compared
Terminal-Bench Hard
Agentic coding and terminal use
42.4%
AA-Omniscience Accuracy
Proportion of correctly answered questions
15.0%
AA-Omniscience Hallucination Rate
Rate of incorrect answers among non-correct responses
16.1%
Last updated Aug 1, 2026
Artificial AnalysisBetter than 89% of models compared
