MiniMax M3.1 Flash
MultimodalMiniMax M3.1 Flash is the fast, low-cost successor in MiniMax's M3 family, released 2026-10-19. It is a multimodal reasoning model with a 1M-token context window, native image and video input, and always-on thinking with five effort levels. Sold at $0.15/$0.60 per million tokens through the MiniMax API, M Plan and MiniMax Code; the open-weight checkpoints are announced but not yet published.
Key Specifications
Parameters
-
Context
1.0M
Release Date
October 19, 2026
Average Score
56.4%
Timeline
Key dates in the model's history
Announcement
October 19, 2026
Last Update
October 20, 2026
Today
October 27, 2026
Technical Specifications
Parameters
-
Training Tokens
-
Knowledge Cutoff
-
Family
-
Capabilities
MultimodalZeroEval
Pricing & Availability
Input (per 1M tokens)
$0.15
Output (per 1M tokens)
$0.60
Max Input Tokens
1.0M
Max Output Tokens
524.3K
Supported Features
Function CallingStructured OutputCode ExecutionWeb SearchBatch InferenceFine-tuning
Benchmark Results
Model performance metrics across various tests and benchmarks
Other Tests
Specialized benchmarks
Agents' Last Exam
Agents' Last Exam; official Overall setting. • Self-reported
AutomationBench v1.0.6
AutomationBench v1.0.6; evaluated on 600 tasks. • Self-reported
BabyVision
BabyVision (w/ tools); Claude Code 2.1.285 harness; image side length 1,500 pixels. • Self-reported
BankerToolBench
OpenCode 1.17.18 harness; MiniMax-M2.7 as judge model. • Self-reported
CharXiv-R
CharXiv RQ; GPT-5.5 as the judge model. • Self-reported
CyberGym
CyberGym; Claude Code 2.1.263 harness; 1M context window. • Self-reported
DeepSWE 1.1
DeepSWE v1.1; mini-SWE-agent 2.4.5 harness; 512K context window. • Self-reported
GDP.pdf
GDP.pdf • Self-reported
Humanity's Last Exam (no tools, text-only)
HLE w/o tools; text-only subset consisting of 2,158 questions. • Self-reported
LOCA-Bench (256k)
LOCA-Bench 256K; official implementation. • Self-reported
NL2Repo
NL2Repo-Bench; Claude Code 2.1.233 harness; 512K context window. • Self-reported
OSWorld 2.0
OSWorld 2.0; 0808 Full release. • Self-reported
PostTrainBench
PostTrain Bench v1.1; Claude Code 2.1.251 harness; 512K context window. • Self-reported
Terminal-Bench 4.0
Terminal-Bench 4.0; Codex 0.149.1 harness. • Self-reported
SWE Refactor Bench
Claude Code 2.1.220 harness; 200K maximum context; otherwise official evaluation setup. • Self-reported
License & Metadata
License
proprietary
Announcement Date
October 19, 2026
Last Updated
October 20, 2026
Compare MiniMax M3.1 Flash
All comparisonsSimilar Models
All ModelsMiniMax M3
MiniMax
MM428.0B
Released:Jun 2026
Price:$0.30/1M tokens
MiniMax M2.5
MiniMax
MM230.0B
Released:Feb 2026
MiniMax M2.7
MiniMax
Released:Mar 2026
Price:$0.30/1M tokens
MiniMax M2.1
MiniMax
Best score:0.9 (TAU)
Released:Dec 2025
Price:$0.30/1M tokens
Sakana Namazu
Sakana AI
MM
Released:Aug 2026
Price:$0.95/1M tokens
GLM-5V-Turbo
Zhipu AI
MM
Released:Apr 2026
Price:$1.20/1M tokens
GPT-5.1 Codex
OpenAI
MM
Released:Nov 2025
Price:$1.25/1M tokens
GPT-6.1 Sol
OpenAI
MM
Released:Sep 2026
Price:$2.00/1M tokens
Recommendations are based on similarity of characteristics: developer organization, multimodality, parameter size, and benchmark performance. Choose a model to compare or go to the full catalog to browse all available AI models.