Research noteResearched 20 Oct 2026 · published 20 Oct 2026, 13:15 IST

Research note: MiniMax launches M3.1 Flash at $0.15 per million input tokens, with open weights planned

How we checked MiniMax M3.1 Flash pricing and separated company benchmarks from independent ones.

Research led by Mayank Sahu

Compiled with AI assistance (Claude by Anthropic)

Read the story →

At a glance

2searches run

3sources read

2sources used

6facts labelled

Limitedevidence strength · 1 of 5 facts (20%)

How we rate evidence: Most facts rest on a single report or on the company’s own claims. We count the share of known facts labelled Confirmed or Reported: 70% or more is Strong, 40–69% Mixed, below 40% Limited.

From sources to story

How many sources we read, how many we used, and how many are official or archived.

Read3Every source we opened

Used2Relied on for the story

Cited2Linked in the story

Official1From the company or a public body

Archived1Saved on the Internet Archive

How sure we are, fact by fact

1131

Confirmed · 1 — An official source says it: lab docs, pricing page, notice.
Company claim · 1 — Only the company says it. Not checked by anyone else yet.
Unconfirmed · 3 — One report only, shown with its source.
Not known · 1 — Not announced or not answered yet.

Evidence map

Each row is a fact from the story; each column is a source. A filled dot means that source supports the fact.

Fact	Label	1OFF	2OUT
M3.1 Flash costs $0.15 input, $0.60 output and $0.015 cached input per million tokens.	Confirmed
MiniMax says agent tasks cost 70% less than on its predecessor.	Company claim
Artificial Analysis gives it 42 on its Intelligence Index, up from 29 for M3.	Unconfirmed
On AA-Omniscience it scored 56% correct with a 93% hallucination rate.	Unconfirmed
Open weights will be released later; VentureBeat reports 23 October.	Unconfirmed
Licence terms for the open weights.	Not known

Question

What we set out to check

The model’s price, availability, benchmark results and open-weight plans.

Findings

What we found

MiniMax’s post and VentureBeat give the same prices.
MiniMax’s benchmark scores are its own; VentureBeat adds Artificial Analysis results.
An earlier “Preview” version from September had different access terms; we did not mix the two.

Discrepancies

Numbers that differ

MiniMax gives CyberGym as 82%; VentureBeat lists 82.2. We do not quote it in the story.

Method

Labels we chose and why

Confirmed for prices on MiniMax’s page.
Claim for MiniMax’s cost-saving figure.
Unconfirmed for Artificial Analysis results and the weights date, which we saw in one outlet.
We explain the hallucination metric so readers do not misread it.

Open questions

What we could not verify

Artificial Analysis’s own pages.
Licence terms for the weights.

Can you help confirm one of these? Send us a source. When we find an answer, we update this note and the story, with the date.

Next steps

How we could improve this

Check Artificial Analysis directly and update the promise tracker on 23 October.

Research timeline

Research20 Oct 2026
Research note published20 Oct 2026, 13:15 IST
Story published20 Oct 2026, 13:25 IST

What we searched

MiniMax M3.1 Flash model
MiniMax M3.1 Flash release $0.15 input price 70% cheaper

Every source we read

Source	Used?	Why
MiniMax blog Archived copy	USED	Prices, speed and company benchmarks.
VentureBeat	USED	Independent Artificial Analysis results and weights date.
DataNorth	NOT USED	About the earlier September preview, not this release.

Sources cited in the story

Research notes show our working. They are not peer-reviewed papers. Data: facts as JSON · How we report