🔍 Read the full analysis: How MiniMax M3.1 Flash Fared Against Open Models In Our Newsroom Test on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
MiniMax released M3.1 Flash on October 19, 2026, pricing it at $0.15 per million input tokens and $0.60 per million output tokens. The supplied review reports a score of 43 on Artificial Analysis’s Intelligence Index and describes a newsroom test of 24 articles, but the source material cuts off before giving that test’s results. The model’s weights are promised for October 23.
MiniMax released M3.1 Flash on October 19, pricing the model at $0.15 per million input tokens and $0.60 per million output tokens as it targets agent-based work. A review published by ThorstenMeyerAI.com compared the launch claims with Artificial Analysis measurements and a 24-article newsroom test; the supplied material reports the test setup but not its scores, leaving that part of the performance comparison unresolved.
MiniMax offers M3.1 Flash through its API, M Plan subscription, MiniMax Code and OpenRouter. OpenRouter listed the model free for its first 72 hours. The listed price for cached input is $0.015 per million tokens. For comparison, OpenRouter lists the previous M3 at $0.30 per million input tokens and $1.20 per million output tokens. MiniMax says the new model costs about 70% less per agent task, a company claim rather than an independently established result in the review.
MiniMax says everyday generation speed exceeds 200 tokens per second, with a median above 170, and describes the model as nearly three times faster than M3. The review cites Artificial Analysis measurements of 120 tokens per second and 1.08 seconds to the first token. Those results differ from MiniMax’s stated speed, though the source does not provide enough detail to reconcile the testing methods. The model has a one-million-token context window; its design details and parameter count have not been officially disclosed.
On Artificial Analysis’s Intelligence Index, M3.1 Flash scored 43. The review lists DeepSeek V4.1 Flash 1017 and GLM-5.3 at 45, MiMo-V2.6-Pro at 46, Step 5 Preview at 44, and Kimi K3 at 44. Artificial Analysis measured M3.1 Flash using 180 million tokens to complete its index, compared with a 99 million-token median. The review’s author cautions that low token prices do not alone establish low overall costs when a model uses more tokens to finish tasks.
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 Versus Measured Performance
M3.1 Flash’s lower per-token price could appeal to teams running agents or generating large volumes of text. But buyers need more than a rate card: speed, output length and accuracy can change the cost and usefulness of a completed task. The independent index results suggest that M3.1 Flash is competitive but does not lead the listed open-weight field on that measure.
The newsroom benchmark was designed to test a different question: whether the model can turn source material into accurate, useful articles without adding unsupported details. That matters for publishers considering models in editorial workflows, where a fluent draft is not enough if it distorts the source. Because the supplied review excerpt omits the scores, it cannot support a conclusion about how M3.1 Flash performed against the other tested models on that task.
Products Worth Considering

LG 27GR83Q-B 27-inch Ultragear QHD IPS Gaming Monitor

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

Bose QuietComfort Ultra Earbuds (2nd Gen), Wireless Noise-Cancelling Earbuds, Black
As an affiliate, we earn on qualifying purchases.
From Space Bunny to M3.1 Flash
Before its public naming, the model operated anonymously on OpenRouter as Space Bunny Alpha. The review says it appeared 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 separately says the model processed more than 200 trillion tokens overall; these figures refer to different scopes and are company-reported.
MiniMax says M3.1 Flash uses “more aggressive sparsity,” but has not supplied a detailed design description or released the weights. It says the weights will arrive on October 23. Until then, Artificial Analysis classifies the model as proprietary. MiniMax’s published benchmark charts, as transcribed in the review, show M3.1 Flash ahead on Automation-Bench, while DeepSeek V4.1 Flash 1017 scores higher on DeepSWE and Terminal-Bench. Those figures are self-reported and have not been independently reproduced in the supplied account.
““more aggressive sparsity””
— MiniMax
Newsroom Scores Are Missing
The source excerpt describes a 24-article newsroom test but ends immediately before the results table. It says 12 news sources were each turned into an English and a German article, with two judges scoring source fidelity, invented facts, clarity, usefulness, structure and polish. The per-model results, including M3.1 Flash’s scores and its comparison with other models, are not present in the supplied material.
Other open questions include whether the promised weights will be released on schedule, what license will govern them, and how the launch-period prompts were handled. The review says that during the Space Bunny period the provider could retain prompts but could not use them for training, and that MiniMax had not explained what happened to the retained data. MiniMax has also not confirmed a parameter count reported by another outlet.
Weight Release and Full Test Results
The next stated milestone is MiniMax’s promised October 23 weight release. Its timing, license and technical specifications will determine whether developers can independently inspect and run the model rather than access it only through hosted services. The supplied account does not confirm that the release has occurred.
For the newsroom comparison, readers would need the missing results table or a later publication of the scores to judge M3.1 Flash against the other tested models. Until those figures are available, the supported conclusions are limited to the release terms, the reported independent index measurements and the test design—not a ranking on newsroom performance.
Key Questions
When did MiniMax release M3.1 Flash?
MiniMax released M3.1 Flash on October 19, 2026, according to the source review.
How much does M3.1 Flash cost?
The listed price is $0.15 per million input tokens, $0.60 per million output tokens and $0.015 per million cached input tokens. OpenRouter offered it free for its first 72 hours.
How did M3.1 Flash score on the independent index?
It scored 43 on Artificial Analysis’s Intelligence Index. The review lists several open-weight models above that score, including MiMo-V2.6-Pro at 46 and DeepSeek V4.1 Flash 1017 and GLM-5.3 at 45.
What did the newsroom test find?
The supplied material describes 24 articles in English and German, judged by GPT-6.1 Sol and Claude Opus 5.5. It does not include the results table, so M3.1 Flash’s newsroom score and ranking cannot be confirmed from this excerpt.
When are the model weights expected?
MiniMax said the weights would arrive on October 23, 2026. The source does not confirm whether they were released.
Source: ThorstenMeyerAI.com
