MiniMax M3.1 Flash In Our Newsroom: How It Stacked Up Against Open Models

🔍 Read the full analysis: MiniMax M3.1 Flash In Our Newsroom: How It Stacked Up Against Open Models on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

MiniMax released M3.1 Flash on October 19, presenting it as a lower-cost, faster model for agent tasks and setting API prices at $0.15 per million input tokens and $0.60 per million output tokens. A review published by ThorstenMeyerAI.com on October 20 found that independent benchmark results put the model behind several open-weight leaders on intelligence and measured speed; the review also describes a newsroom test but does not provide its scores in the supplied material.

MiniMax made M3.1 Flash available through its API, M Plan subscription, MiniMax Code and OpenRouter. OpenRouter offered access free for the first 72 hours. The listed cached-input price is $0.015 per million tokens. The review says those per-token rates are half the listed prices of its predecessor, M3. MiniMax also claims the model costs about 70% less per agent task, a company figure that is not independently verified in the material provided.

MiniMax advertises a one-million-token context window and adjustable reasoning effort. It says everyday output speed exceeds 200 tokens per second, with a median above 170, and describes the model as “nearly three times faster” than M3. Artificial Analysis measured 120 tokens per second and 1.08 seconds to the first token. That is a different result from the company’s claim, using the independent index’s testing conditions.

The review reports a score of 43 on Artificial Analysis’ Intelligence Index. It places MiMo-V2.6-Pro at 46, DeepSeek V4.1 Flash 1017 and GLM-5.3 at 45, and Step 5 Preview and Kimi K3 at 44. In MiniMax’s own published benchmark charts, M3.1 Flash leads DeepSeek on Automation-Bench by 1.7 points, while DeepSeek scores higher on DeepSWE and Terminal-Bench. These are company-selected comparisons and self-reported results, not independent reproductions.

Price Versus Measured Performance

M3.1 Flash’s launch matters to developers weighing inference cost against task quality and speed. Its listed token prices are low, and the million-token context window may suit workflows that process long inputs. But price per token does not by itself show the cost of completing a task: the review says Artificial Analysis recorded 180 million tokens used during its index evaluation, compared with a median of 99 million. The source offers no broader cost-per-successful-task comparison, so that observation should not be treated as a general estimate for every use.

The gap between MiniMax’s speed claim and the index’s measurement also gives buyers reason to test the model on their own workloads. For newsroom or other fact-sensitive uses, benchmark scores are only part of the picture. The review says it tested 24 articles from 12 sources in English and German, scored by two independent judges, but the supplied account ends before giving the model’s results. Its real-world editorial performance therefore cannot be established from this material.

From Space Bunny to M3.1 Flash

Before its public identification, the model ran anonymously on OpenRouter under the name Space Bunny Alpha. According to the review, it appeared on September 23, ranked as OpenRouter’s most-used model for 10 days and processed 56.7 trillion tokens on that platform before being removed on October 5. MiniMax says the model processed more than 200 trillion tokens overall, a company-reported figure.

The review says that during the anonymous period the provider could retain prompts but could not use them for training. It reports that MiniMax has not explained what happened to that data. The weights were not available at the time of publication; MiniMax said they would arrive October 23. Until a release and license are available, the review describes M3.1 Flash as proprietary, and neither its license nor official parameter count is confirmed.

MiniMax’s benchmark chart compares M3.1 Flash with several models, but the review notes that some prominent competitors were represented by older versions. It also says a claim that M3.1 Flash ranked first among Chinese models on Agents’ Last Exam needs qualification: MiniMax’s chart omits DeepSeek from that row, while DeepSeek’s model card reports a score of 35.4 under a potentially different setup. The figures are not directly comparable without aligned test conditions.

““nearly three times faster” than M3”

— MiniMax

Weights, Privacy and Test Results

Several points remain unresolved in the source material. The promised October 23 weights release had not yet occurred as of the review’s October 20 publication, and no license or official parameter count was available. A separate outlet reportedly gave a parameter estimate, but the reviewer says MiniMax had not confirmed it and excludes it from the analysis.

The review does not disclose what became of prompts retained during the Space Bunny Alpha period. It also does not reproduce MiniMax’s benchmark results independently. Most importantly for the stated newsroom comparison, the source describes the test design—12 news sources, each produced in English and German, judged by GPT-6.1 Sol and Claude Opus 5.5—but the supplied material contains no score table or findings. No conclusion about how M3.1 Flash ranked in that test is supported here.

Watch for the Weights Release

The next stated milestone is MiniMax’s planned October 23 release of the weights. If released, the weights and license details would clarify whether developers can run or adapt the model outside MiniMax’s hosted services. The source does not confirm the release in advance, so its timing and terms remain subject to verification.

Further evaluation would help resolve the gaps between the company’s claims and independent measurements. That includes reproducing benchmark results, testing speed and token use under comparable conditions, and publishing the missing newsroom scores. Until then, buyers can confirm the listed API prices and availability, but should treat claimed agent-task savings and unreported editorial outcomes as unverified.

Key Questions

When did MiniMax release M3.1 Flash?

MiniMax released M3.1 Flash on October 19, 2026, according to the October 20 review. It was available through the MiniMax API, M Plan, MiniMax Code and OpenRouter.

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 there.

Did independent tests confirm MiniMax’s speed claim?

No. MiniMax claimed more than 200 tokens per second in everyday use, while Artificial Analysis measured 120 tokens per second. The figures come from different sources and testing conditions; the review does not establish a direct, controlled comparison.

How did M3.1 Flash perform in the newsroom test?

The review describes a test of 24 articles based on 12 news sources, produced in English and German and evaluated by two judges. The supplied source material does not include the scores or conclusions, so its newsroom performance cannot be determined from the information provided.

Are M3.1 Flash’s weights available?

They were not available at the time of the October 20 review. MiniMax said the weights would arrive October 23, but the material does not confirm whether that release happened or state what license would apply.

Source: ThorstenMeyerAI.com