Editorial illustration for MiniMax M3.1 Flash and Deepseek V4.1 split China's AI pricing battle
MiniMax M3.1 Flash Cuts AI Costs in Half
MiniMax M3.1 Flash and Deepseek V4.1 split China's AI pricing battle
MiniMax put a number on how cheap reasoning can get. The company's new M3.1 Flash model scores 42 points on the Artificial Analysis Intelligence Index, a 13-point jump over its predecessor M3, while costing less than half as much per task to run. At roughly $0.22 per task, with API pricing set at $0.15 per million input tokens and $0.60 per million output tokens, the model lands in the same bracket as Zhipu's GLM-5.3-Flash and ahead of Google's Gemini 3.8 Flash, though Deepseek's V4.1 Flash 1017 still leads the pack at 45 points.
The real story isn't just the price tag, it's what MiniMax gave up to get there. The model burns through far fewer output tokens than its competitors and runs three times faster than M3, according to the company. It handles up to a million tokens of context and takes both text and images as input. But speed and cost cuts came with a tradeoff on reliability, one that shows up clearly in how the model behaves when it doesn't actually know the answer.
MiniMax has released M3.1 Flash, a new reasoning model that scores 42 points on the Artificial Analysis Intelligence Index. That's 13 points higher than its predecessor M3, at less than half the cost per task.
Why this matters
For developers choosing between MiniMax M3.1 Flash and Deepseek V4.1 Flash 1017, there's no clean winner, and that's the point worth sitting with. Deepseek takes the Intelligence Index and CritPt, MiniMax claws back ground on complex reasoning and cost. If you're building a product on a budget, a 13-point jump at half the price per task matters more than a three-point gap on a leaderboard nobody's customers read.
We've watched Chinese labs treat price as the main lever for months now, and M3.1 Flash confirms the floor keeps dropping rather than settling. For founders, this means the calculus shifts from "which model is smartest" to "which model is smart enough at a price that doesn't wreck margins." For researchers, the GLM-5.3-Flash tie and the Gemini 3.8 Flash edge-out suggest the mid-tier reasoning bracket is getting crowded fast, with gains measured in single digits rather than leaps. Watch whether Deepseek answers with its own price cut, because that's the move that actually decides who wins this round.
Related Reading
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google to End Free Access to Gemini AI for Non-Subscribers
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
From Archives Replika founder Kuyda launches Wabi, multiplayer AI messenger app
From Archives Anthropic, Google block OpenCode API keys after hackers ran up user charges
Common Questions Answered
How does MiniMax M3.1 Flash's performance compare to its predecessor M3?
MiniMax M3.1 Flash scores 42 points on the Artificial Analysis Intelligence Index, representing a 13-point improvement over its predecessor M3. Notably, this significant performance jump was achieved while reducing the cost per task to less than half of what M3 required.
What is the API pricing structure for MiniMax M3.1 Flash?
MiniMax M3.1 Flash is priced at $0.15 per million input tokens and $0.60 per million output tokens, with an approximate cost of $0.22 per task. This pricing places it in the same competitive bracket as Zhipu's GLM-5.3-Flash and ahead of Google's Gemini 3.8 Flash.
How does MiniMax M3.1 Flash compare to Deepseek V4.1 Flash 1017 for developers?
While Deepseek V4.1 Flash 1017 leads on the Intelligence Index and CritPt benchmarks, MiniMax M3.1 Flash offers advantages in complex reasoning capabilities and significantly lower costs. For budget-conscious developers, MiniMax's 13-point performance improvement at half the price per task may be more valuable than Deepseek's marginal three-point leaderboard advantage.
What does the Artificial Analysis Intelligence Index measure for AI models?
The Artificial Analysis Intelligence Index is a benchmarking metric used to evaluate and compare the performance of AI reasoning models like MiniMax M3.1 Flash and Deepseek V4.1 Flash 1017. It provides a standardized scoring system that helps developers assess model capabilities across different providers.
Further Reading
- MiniMax M3.1 Flash shows China's AI price race has no floor in sight - The Decoder
- MiniMax's new M3.1 Flash beats DeepSeek and Opus 5 on a key business benchmark at lower cost — but with a high hallucination rate - VentureBeat
- MiniMax-M3.1-Flash Intelligence, Performance & Price Analysis - Artificial Analysis
- MiniMax M3.1 Flash Scores 42 and Costs $0.22 a Task - Orca Router
