All AI models · LLM benchmarks · Methodology

MiniMax

MiniMax-M3

MiniMax multimodal model for long-context coding, perception, and agent planning

Model facts

Context window
1.05M tokens
Maximum output
512K tokens
Cheapest paid input
$0.225 per 1M tokens
Cheapest paid output
$0.9 per 1M tokens
Open weights
yes
Providers
51

Capabilities and modalities

reasoning, tool calling, structured output, attachments, temperature control, open weights, text, image, video, pdf.

Local hardware estimate

At 4K context: Q4_K_M 262.3 GiB, Q8_0 461.2 GiB, F16 834.1 GiB working memory.

Parameters
427.0 billion
Architecture
minimaxm3sparse
Model license
other

Planning estimate adapted from the MIT-licensed whichllm estimator using metadata from Hugging Face; not a vendor minimum requirement.

Artificial Analysis benchmarks

Compare this model on the full LLM benchmark leaderboard.

ConfigurationAAICodingMathOutput tokens/s
default45.458.6not measured77.1

Observed price history

  • 2026-08-24: $0.225 input / $0.9 output per 1M tokens

API providers