MiniMax M3.1 Flash Pricing
Per-million-token rates and the cached-input discount for MiniMax's Flash-tier model, with a worked example.
MiniMax M3.1 Flash pricing is $0.15 per million input tokens and $0.60 per million output tokens. Cached input tokens cost $0.015 per million on the MiniMax application programming interface (API). The Chinese artificial intelligence (AI) lab MiniMax announced the next-generation Flash model on October 19, 2026, listing it under the model identifier MiniMax-M3.1-Flash.
MiniMax M3.1 Flash input and output rates are half the rates listed for MiniMax M3. The announcement does not publish M Plan prices.
MiniMax M3.1 Flash Price per Million Tokens
MiniMax M3.1 Flash costs $0.15 per million uncached input tokens, $0.015 per million cached input tokens, and $0.60 per million output tokens. MiniMax published these flat rates on October 19, 2026, in its announcement.
Prompt caching reduces input expense by 90 percent compared to uncached input. When repeated context hits the cache, MiniMax bills $0.015 per million tokens instead of the standard $0.15 rate. Output generation remains billed at the flat rate of $0.60 per million tokens.
| Token Class | Price per 1 Million Tokens | Price per 1 Thousand Tokens |
|---|---|---|
| Uncached Input | $0.15 | $0.00015 |
| Cached Input | $0.015 | $0.000015 |
| Output | $0.60 | $0.00060 |
Worked Calculation for Mixed API Workloads
A standard API workload processing 10 million input tokens and 2 million output tokens costs $2.70 with zero caching, or $1.35 if all input hits prompt cache.
For an uncached run of 10 million input tokens, multiplying 10 by $0.15 yields $1.50 in input expense. Adding 2 million output tokens at $0.60 per million yields $1.20 in generation expense, producing a $2.70 total. If an application caches the entire 10 million input tokens across recurring agent turns, 10 multiplied by $0.015 equals $0.15 for input, bringing the total to $1.35.
Developers comparing token economics across providers can review our AI Model Pricing Index to contrast MiniMax rates with other models such as DeepSeek V4.1 Flash.
- 10M uncached input tokens: 10 x $0.15 = $1.50
- 10M cached input tokens: 10 x $0.015 = $0.15
- 2M output tokens: 2 x $0.60 = $1.20
- Total cost without caching: $1.50 + $1.20 = $2.70
- Total cost with complete input caching: $0.15 + $1.20 = $1.35
Comparison With MiniMax M3 Pricing
MiniMax M3.1 Flash input and output rates are half of listed MiniMax M3 pay-as-you-go rates. On the MiniMax pay-as-you-go pricing page, MiniMax lists MiniMax M3 at $0.30 per million input tokens, $0.06 per million cached input tokens, and $1.20 per million output tokens for contexts up to 512,000 tokens.
Comparing these listed numbers shows that MiniMax M3.1 Flash uncached input at $0.15 is half of the $0.30 M3 rate, while output at $0.60 is half of the $1.20 M3 rate. Cached input drops from $0.06 on M3 to $0.015 on M3.1 Flash, representing a 75 percent reduction.
| Model | Input per 1M | Cached Input per 1M | Output per 1M |
|---|---|---|---|
| MiniMax M3.1 Flash | $0.15 | $0.015 | $0.60 |
| MiniMax M3 (up to 512k context) | $0.30 | $0.06 | $1.20 |
| Difference | 50% lower | 75% lower | 50% lower |
In its launch announcement, MiniMax claims that usage cost for agentic tasks is 70 percent lower than its predecessor, though MiniMax does not name which model served as that predecessor. MiniMax also claims a generational leap in agentic tasks, coding, Computer Use, and cybersecurity, stating online inference runs nearly three times faster. An earlier open-weight model from MiniMax is detailed in our guide to MiniMax M2.
Access Channels and Deployment Options
Developers can access MiniMax M3.1 Flash through three distinct product surfaces announced by MiniMax. MiniMax offers the model through the MiniMax API on platform.minimax.io.
The second is the MiniMax M Plan. MiniMax Code, a desktop app, is the third. MiniMax states that model weights will be released later, though no specific release date or license terms have been published.
MiniMax also disclosed that a Pro model that pushes intelligence further will arrive later.
- MiniMax API: platform.minimax.io
- MiniMax M Plan: access via minimax.io/m-plan
- MiniMax Code: desktop software application
- Weights: promised for later, no date published
Unpublished Specifications and Verification Steps
MiniMax leaves several operational specifications unpublished in its launch documentation. The announcement does not state the maximum context window, total parameter count, weights license, organization rate limits, or free tier provisions.
Pricing terms for the MiniMax M Plan are also omitted from the announcement. Specific billing rules regarding whether tool calls or internal reasoning tokens bill at standard output rates are not published in the announcement.
Before scheduling production traffic, confirm current per-token rates on the MiniMax pay-as-you-go pricing page to verify current MiniMax M3.1 Flash pricing.
Frequently Asked Questions
- MiniMax M3.1 Flash costs $0.15 per million uncached input tokens, $0.015 per million cached input tokens, and $0.60 per million output tokens on the MiniMax API.
- Cached input tokens cost $0.015 per million tokens, providing a 90 percent discount compared to the $0.15 per million uncached input rate.
- Yes. MiniMax M3.1 Flash input at $0.15 and output at $0.60 are exactly half of the $0.30 input and $1.20 output rates listed for MiniMax M3 on the pay-as-you-go pricing table.
- MiniMax has not published details regarding a free tier in the launch announcement.
Planning to Use MiniMax M3.1 Flash?
Book a free AI workflow audit with Layer3Labs to map where a model like this fits your team's work.
Book a Consultation