🔍 Read the full analysis: Can MiniMax M3.1 Flash Outperform Open Models In A Newsroom Test? on ThorstenMeyerAI.com
Get hardware and tech essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
MiniMax released M3.1 Flash on October 19, 2026, at $0.15 per million input tokens and $0.60 per million output tokens. The source review reports mixed benchmark results: the model scored 43 on Artificial Analysis’ Intelligence Index, while its planned 24-article newsroom test is described but its scores are not provided in the available material. Its weights are scheduled for October 23, and several performance and data-handling questions remain open.
MiniMax released M3.1 Flash on October 19, pricing it at $0.15 per million input tokens and $0.60 per million output tokens. A review by ThorstenMeyerAI.com compared the model’s claims with an independent index and a newsroom evaluation, but the supplied account does not include the newsroom scores, so it cannot establish whether Flash outperformed open-weight alternatives on that test.
MiniMax describes M3.1 Flash as a low-cost, fast model for agent work. Its listed API price is half the per-token rate of its predecessor, M3, according to the review: M3 costs $0.30 per million input tokens and $1.20 per million output tokens. MiniMax also claims the new model is about 70% cheaper per agent task. That comparison is the company’s claim, not an independently verified cost result.
MiniMax says everyday generation speed exceeds 200 tokens per second, with a median above 170, and calls the model nearly three times faster than M3. In contrast, Artificial Analysis measured 120 tokens per second and 1.08 seconds to the first token. The index gave M3.1 Flash an Intelligence Index score of 43, below DeepSeek V4.1 Flash 1017 and GLM-5.3, both at 45, in the comparison reported by the review.
The newsroom evaluation used 12 real news sources in English and German, producing 24 articles. Two judges, GPT-6.1 Sol and Claude Opus 5.5, each rated articles on source fidelity, invented facts, clarity, usefulness, structure and polish. However, the source material supplied here ends before the results table: it names the test setup but provides no model scores or comparisons from that test.
Infographic 01 · MiniMax M3.1 Flash
A step behind on the index. Ahead of every open model in our newsroom.
An independent intelligence index puts MiniMax M3.1 Flash a step behind the open-weight leaders. Writing real news from real sources, it beat all of them.
Independent index
Artificial Analysis Intelligence Index, 10 evaluations. Bar length on a 0–50 scale.
Artificial Analysis lists M3.1 Flash as proprietary until its weights are released.
Our newsroom test
Mean score, English articles, 1–5 by GPT-6.1 Sol and Claude Opus 5.5 with the source in context. Bar length on a 0–5 scale.
In German, M3.1 Flash (3.36) also leads every open model, but GPT-6 Luna is clearly ahead (3.62).
Sources: Artificial Analysis, MiniMax-M3.1-Flash and open-weights models (20 Oct 2026); Thorsten Meyer AI newsroom benchmark, 12 sources × English + German, run 20 Oct 2026.
Powered by Thorsten Meyer AI
Price and Performance Trade-Off
The release puts a lower-priced model into competition for news production and agent tasks, where token costs and factual reliability both matter. The price could make repeated drafting or summarization less expensive, but a low rate per token does not by itself show that a model is cheaper for a complete task. Artificial Analysis found M3.1 Flash used 180 million tokens during its index evaluation, compared with a median of 99 million. That is a raw comparison from the reported evaluation, not proof of how much it would cost to produce a newsroom article.
For publishers, the central question is whether the model can follow source material without adding unsupported details. The described newsroom test was designed to assess that problem, but the missing score table prevents readers from judging the result or comparing Flash with open models. The available evidence supports a mixed picture on price, speed and index performance, not a conclusion that it wins newsroom work.
As an affiliate, we earn on qualifying purchases.
From Space Bunny to M3.1 Flash
Before the announcement, the model appeared anonymously on OpenRouter under the name Space Bunny Alpha. According to the review, it went live on September 23, led OpenRouter’s usage ranking for 10 days, and processed 56.7 trillion tokens on the platform before being removed on October 5. MiniMax says total usage across services exceeded 200 trillion tokens. Those figures are reported by the review and MiniMax, respectively.
MiniMax published its own benchmark comparisons, but the review notes that the company selected the rival models and compared Flash with some older releases. Its charts show Flash ahead on Automation-Bench, while DeepSeek V4.1 Flash 1017 scored higher on DeepSWE and Terminal-Bench. These are company-reported results; the review says they have not been independently reproduced. Artificial Analysis treats M3.1 Flash as proprietary while its weights remain unavailable.
MiniMax says the model uses “more aggressive sparsity,” but has not published a parameter count in the material reviewed. The weights were promised for October 23; no license was yet available. The review says an outlet reported a parameter estimate, but MiniMax had not confirmed it, so the review excluded that figure.
“More than 200 tokens per second in everyday use and a median above 170.”
— MiniMax
As an affiliate, we earn on qualifying purchases.
Missing Test Scores and Data Details
The supplied review material does not include the results of the 24-article newsroom test, despite introducing a results table. It is therefore not possible to say from this material whether M3.1 Flash outperformed open models on article accuracy, source fidelity or writing quality. The judges and scoring criteria are described, but comparative scores and any significance testing are absent.
Other points remain unsettled. MiniMax’s benchmark and speed claims have not been independently reproduced in the account, the model’s parameter count and license are undisclosed, and the promised weights had not yet arrived. MiniMax also had not explained what happened to prompts submitted while Space Bunny Alpha was available. The review says the provider could retain prompts during that period, though it was not allowed to train on them; the status of that retained data is unclear.
AI newsroom content creation software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Weights and Results to Watch
The stated next milestone is the planned October 23 weight release. Its arrival, license terms and parameter details could clarify how the model can be inspected and deployed outside MiniMax’s services. The release date is a company commitment reported by the review, not confirmation that the weights have since become available.
For a newsroom comparison, the missing scores are the immediate gap: the test’s English and German results, along with the comparator models’ scores and judging method, are needed to support any claim that Flash outperformed open alternatives. Readers should also look for independent replication of benchmark and speed figures and a clear statement from MiniMax about prompts retained during the anonymous OpenRouter period.
As an affiliate, we earn on qualifying purchases.
Key Questions
When did MiniMax release M3.1 Flash?
MiniMax announced and released M3.1 Flash on October 19, 2026, according to the supplied review.
How much does M3.1 Flash cost?
The listed rate is $0.15 per million input tokens and $0.60 per million output tokens. Cached input is listed at $0.015 per million tokens.
Did M3.1 Flash beat open models in the newsroom test?
The supplied source describes a 24-article test but does not include its scores. It does not provide enough information to determine whether M3.1 Flash outperformed open-weight models in that evaluation.
How did it score on the independent index?
Artificial Analysis gave M3.1 Flash an Intelligence Index score of 43 and measured output speed at 120 tokens per second. The review’s comparison lists DeepSeek V4.1 Flash 1017 and GLM-5.3 at 45.
Were the model weights available at release?
No. The review said the weights had not yet been released and that MiniMax had scheduled them for October 23. The supplied material does not confirm whether that release occurred.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
