# MiniMax AI model releases

> Developing situation record from CLSTR: https://clstr.news/situations/minimax-ai-model-releases
> Updated: 2026-10-19T18:10:30.000Z. Sources: 17. Developments: 2.

AI developer MiniMax has released a series of specialized models focused on developer productivity and reasoning efficiency.

In late September 2026, the company launched MiniMax-M3.1-Flash-Preview. This model was integrated into the MiniMax Code environment to assist with programming tasks such as bug localization and testing. It featured a 1-million-token context window and supported multimodal inputs including text, images, and video.

By mid-October 2026, MiniMax released the M3.1 Flash reasoning model. This iteration aimed to reduce the costs of running AI agents by 70% compared to the previous M3 model, while increasing inference speeds up to 200 tokens per second. While the model outperformed competitors like Gemini 3.8 Flash and GLM-5.3-Flash in certain benchmarks, testing indicated a “high hallucination rate of 93%” because the model tended to attempt all questions rather than declining uncertain ones.

On October 19, 2026, further details emerged regarding the M3.1 Flash model, which utilizes a mixture-of-experts (MoE) architecture. Previously known under the anonymous pseudonym ‘Space Bunny’, the model features 428 billion total parameters. It has demonstrated strong performance in benchmarks for cybersecurity (CyberGym) and computer operation (OSWorld V2). MiniMax has positioned the model with aggressive pricing at $0.15 per million input tokens and $0.60 per million output tokens, and announced plans to release the model weights as open weights on October 23.

## Claims

- MiniMax M3.1 Flash has 428 billion total parameters. (reported by 5 sources)
- MiniMax M3.1 Flash reduces AI agent task costs by 70% compared to previous models. (reported by 4 sources)
- DeepSeek open-sourced the DeepInfer inference framework under the Apache 2.0 license. (reported by 4 sources)
- MiniMax M3.1 Flash features a context window of one million tokens. (reported by 4 sources)
- MiniMax M3.1 Flash is priced at $0.15 per million input tokens. (reported by 4 sources)
- DeepInfer is optimized for NVIDIA SM100 GPUs and Huawei Ascend 950 NPUs. (reported by 3 sources)
- MiniMax M3.1 Flash achieved a score of 76.2 on the BankerToolBench evaluation. (reported by 3 sources)
- DeepSeek V4.1-Flash-1017 resolved 77% of tasks on the DeepSWE v1.1 test. (reported by 2 sources)

## Timeline

### 2026-10-19: DeepSeek and MiniMax launch low-cost, high-performance AI models

Chinese AI firms DeepSeek and MiniMax have released new high-speed, low-cost models, M3.1 Flash and V4.1-Flash, to drive efficiency in AI agent and coding tasks.

15 sources. https://clstr.news/cluster/minimax-launches-m31-flash-reasoning-model

### 2026-09-27: MiniMax launches M3. 1-Flash-Preview AI model for developers

MiniMax has released the M3. 1-Flash-Preview, a high-speed AI model for developers featuring a 1-million-token context window and multimodal capabilities for coding and reasoning tasks.

2 sources. https://clstr.news/cluster/minimax-launches-m3-1-flash-preview-ai-model-for-developers

---
Cite as: MiniMax AI model releases. CLSTR, https://clstr.news/situations/minimax-ai-model-releases