🔍 Read the full analysis: MiniMax M3.1 Flash: Comparing Rankings With Hands-On Newsroom Results on ThorstenMeyerAI.com
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. An independent index scored it below several open-weight rivals, while a newsroom test of 24 articles was designed to measure source fidelity and writing quality; the supplied source does not include that test’s results.
MiniMax released M3.1 Flash on October 19, a low-priced model promoted for fast agent tasks, and a review by ThorstenMeyerAI.com compared its claims with an independent index and a newsroom writing test. The reported index score was 43, behind several open-weight competitors; the supplied review text describes the 24-article test setup but does not provide its results.
MiniMax lists M3.1 Flash at $0.15 per million input tokens and $0.60 per million output tokens, half the listed per-token price of M3. It is available through the MiniMax API, M Plan subscription, MiniMax Code and OpenRouter, which offered free access for the first 72 hours. MiniMax also lists a one-million-token context window and adjustable reasoning effort. OpenRouter permits up to 524,288 output tokens.
MiniMax says the model exceeds 200 tokens per second in everyday use, with a median above 170, and calls it nearly three times faster than M3. Artificial Analysis measured 120 tokens per second and 1.08 seconds to first token. These are different measurement settings; the independent result is lower than the company’s stated speed. MiniMax attributes the model’s design to “more aggressive sparsity,” but has not published further architecture details in the material reviewed.
The review also assessed performance on the Artificial Analysis Intelligence Index and a newsroom benchmark using 12 source items in both English and German, for 24 articles total. Two judges scored each article for source fidelity, invented facts, clarity, usefulness, structure and polish. The text provided explains the method and names the judges, but ends before giving the scores, so no newsroom ranking or quality conclusion can be reported.
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 low token prices make it relevant to developers and publishers running repeated model tasks, but price alone does not establish the cost of a completed job. Artificial Analysis found the model used 180 million tokens to complete its index, compared with a 99 million-token median. That observation suggests that token consumption can affect overall economics, though it does not establish the cost or efficiency of every user’s workload.
The difference between MiniMax’s speed claim and the independent measurement also illustrates why model comparisons depend on test conditions. And for newsroom use, a strong general benchmark score would not answer the key question: whether the model stays faithful to supplied reporting and avoids fabricating details. The described editorial test targets that practical issue, but its missing results prevent readers from judging the outcome.
As an affiliate, we earn on qualifying purchases.
From Space Bunny to M3.1
Before the public release, the model appeared anonymously on OpenRouter as Space Bunny Alpha, beginning September 23. The review says it led the platform’s usage ranking for 10 days, processed 56.7 trillion tokens there, and was removed on October 5. MiniMax says total usage across services exceeded 200 trillion tokens. These usage figures are attributed to the review or company, rather than independent measures of model quality.
MiniMax has promised to release the weights on October 23. Until then, Artificial Analysis lists M3.1 Flash as proprietary. The source says MiniMax has not announced a license or official parameter count. A reported parameter estimate of 428 billion total and 23 billion active was not confirmed by MiniMax and is not treated here as established fact.
MiniMax’s own benchmark charts show mixed results against selected rivals. M3.1 Flash scored 62.2 on Automation-Bench, ahead of DeepSeek V4.1 Flash 1017 at 60.5. DeepSeek scored higher on DeepSWE, 77.0 to 73.7, and Terminal-Bench, 42.3 to 38.1. The comparison used models chosen by MiniMax, and some newer competitors were not included, according to the review.
““More aggressive sparsity.””
— MiniMax
As an affiliate, we earn on qualifying purchases.
Newsroom Scores Not Supplied
The source material does not include the newsroom benchmark’s numerical scores or findings. It therefore does not establish how M3.1 Flash compared with other models on article accuracy, invented facts or writing quality. The described test used a free OpenRouter endpoint and a 32,000-token output limit; readers would need the full results to assess the model’s performance under that setup.
Other open questions include the model’s final license and parameter count, the conditions behind MiniMax’s speed claims, and whether its internal benchmark results can be reproduced independently. The review also says that during the anonymous Space Bunny period, the provider could retain prompts but could not use them for training. MiniMax had not clarified in the source what happens to that retained data.
As an affiliate, we earn on qualifying purchases.
Weights and Full Test Results
The next stated milestone is MiniMax’s promised October 23 weight release. A published license and model documentation would clarify how the model can be used and what architecture details are supported. Independent evaluations after release could test the company’s benchmark and speed claims under publicly described conditions.
For the newsroom comparison, the missing score table is needed before drawing conclusions about editorial suitability. Until those results are available, the confirmed picture is limited: M3.1 Flash is released at low listed token prices, its independent index score trails several listed open-weight models, and the supplied material does not reveal how it performed in the article-writing test.
AI model performance testing tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
When was MiniMax M3.1 Flash released?
MiniMax released it on October 19, 2026, through its API, M Plan, MiniMax Code and OpenRouter, according to the review.
What does M3.1 Flash cost?
The listed price 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 the model beat other open-weight models?
Not on the cited Artificial Analysis Intelligence Index: M3.1 Flash scored 43, while listed rivals scored as high as 46. MiniMax’s own benchmark table showed a mixed picture, with different models leading different tests.
What did the newsroom test find?
The supplied source describes a test of 24 articles, made from 12 sources in English and German, scored by two judges. It does not include the scores or conclusions, so the test’s results cannot be stated.
Are M3.1 Flash’s weights available?
Not in the source’s account as of the October 20 review. MiniMax said the weights would arrive on October 23; the license and official parameter count were not yet disclosed.
Source: ThorstenMeyerAI.com
