MiniMax M3.1 Flash is a MiniMax model with 1M context, $0.15 in / $0.60 out per million tokens, last verified October 25, 2026. Open weights: yes. Public-board scores are linked to the publisher; vendor-only figures stay in the watch-out, not on the board.
MiniMax’s final M3.1 Flash multimodal MoE checkpoint and API, with approximately 428B total/23B active parameters and selectable reasoning depth.
Recorded profile verification October 25, 2026·Individual provider fields may change
Pricing tiers: MiniMax PAYG USD/MTok input/output $0.15/$0.60, cached input $0.015. The checked tariff does not establish a separate cache-write fee, TTL or context-band surcharge; unknown is not free. M3.1 Flash does not accept service_tier, so M3’s Priority 1.5x multiplier does not apply. Subscription Credits/M Plan and PAYG are separate billing contracts. Hosted context is 1,000,000; native config has 1,048,576 positions. A hard output cap is not established.
Best for
Multimodal coding agents
Long-context PAYG workflows
Evaluating the exact M3.1 Flash checkpoint
Watch out
Always thinks: disabling or effort=none fails; low/medium/high/xhigh/max defaults max. OpenAI uses reasoning_effort; Anthropic uses output_config.effort. Preserve complete reasoning/tool history. API sampling controls can be ignored, while native inference recommends temperature=1/top_p=0.95. Commercial weight use requires Built with MiniMax M3.1-Flash attribution; above US$20M yearly revenue from the relevant products/services prior written authorization is required, otherwise a one-time notice. Retain notices and comply with prohibited-use terms. Publisher DeepSWE and other benchmark claims are not independent matched observations; none are inherited from M3 or Preview.
History
MiniMax M3.1 Flash timeline
Release and status events for this model that passed two-source verification — dated, sourced, and linked.
MiniMax M3.1 FlashReleaseOctober 19, 2026
MiniMax's dated model log announces M3.1 Flash, the final multimodal model now listed in its API and PAYG documentation.
Answered from the verified figures on this page rather than general guidance.
How much does MiniMax M3.1 Flash cost per million tokens?
MiniMax M3.1 Flash lists base rates of $0.15 per million input tokens and $0.60 per million output tokens. Estimate both input and output token volume and any additional charges for your workload. MiniMax PAYG USD/MTok input/output $0.15/$0.60, cached input $0.015. The checked tariff does not establish a separate cache-write fee, TTL or context-band surcharge; unknown is not free. M3.1 Flash does not accept service_tier, so M3’s Priority 1.5x multiplier does not apply. Subscription Credits/M Plan and PAYG are separate billing contracts. Hosted context is 1,000,000; native config has 1,048,576 positions. A hard output cap is not established.
What is MiniMax M3.1 Flash's context window?
MiniMax M3.1 Flash accepts about 1M tokens of context. That only matters if you routinely send very long documents, large codebases, or multi-turn histories that approach that limit.
What is MiniMax M3.1 Flash best for?
MiniMax M3.1 Flash is a balanced tier from MiniMax. It suits multimodal coding agents, long-context payg workflows, evaluating the exact m3.1 flash checkpoint. Always thinks: disabling or effort=none fails; low/medium/high/xhigh/max defaults max. OpenAI uses reasoning_effort; Anthropic uses output_config.effort. Preserve complete reasoning/tool history. API sampling controls can be ignored, while native inference recommends temperature=1/top_p=0.95. Commercial weight use requires Built with MiniMax M3.1-Flash attribution; above US$20M yearly revenue from the relevant products/services prior written authorization is required, otherwise a one-time notice. Retain notices and comply with prohibited-use terms. Publisher DeepSWE and other benchmark claims are not independent matched observations; none are inherited from M3 or Preview.
Can I self-host MiniMax M3.1 Flash?
MiniMax M3.1 Flash publishes open weights, but self-hosting depends on the licence, hardware footprint, quantisation quality, and serving stack. A hosted API is often cheaper until you have measured throughput and concurrency on your own hardware.
