MiniMax
Low-cost agentic coding models, video and speech via API and M Plan
MiniMax is an AI assistants tool for Developers and Small businesses, offered as freemium ($22/mo). Founded in 2022. Popular alternatives include Claude, DeepSeek, Cursor.
About MiniMax
MiniMax is an AI foundation model company founded, by its own account, in early 2022. It builds its own models for text, video, speech, image and music and sells them through an Open API Platform, the MiniMax Code agent app and an M Plan subscription. For developers the core job is cheap, fast language models for coding and agent work, served behind endpoints that copy the Anthropic and OpenAI formats.
The current text models are MiniMax-M3.1-Flash and MiniMax-M3, both with a 1M-token context window and multimodal input; M3.1 Flash adds an effort setting from low to max. Next to them sit H3 video, speech-2.8 TTS with voice cloning, speech recognition and image-01. Pay-as-you-go API pricing is per million tokens, with a Priority tier at 1.5x. M Plan, aimed at individual developers, costs $22, $55 or $132 a month for Go, Explore and Build, with two months free on yearly billing, and gives a key that works in Claude Code, Codex, Cursor and OpenCode.
It fits developers and teams that run long agent loops and want token costs far below premium models. DeepSeek is the closest low-cost rival, while Claude costs much more per token but brings richer apps and enterprise controls. M Plan is not meant for production and has 5-hour and weekly limits, M2.5 and M2.7 retire on November 18, 2026, and benchmark figures come from MiniMax itself.
MiniMax released M3.1 Flash on October 19, 2026 and launched it on Product Hunt on October 26, 2026. It claims about 3x faster inference, 70% lower agentic task cost and 73.7% on DeepSWE v1.1. Tip: test with a pay-as-you-go key before subscribing.
Key features of MiniMax
- MiniMax-M3.1-Flash at $0.15 input and $0.60 output per 1M tokens
- 1M-token context window on M3.1 Flash and M3
- Effort setting on M3.1 Flash, from low to max, to trade thinking depth for cost
- Anthropic- and OpenAI-compatible endpoints, plus the Responses API
- M Plan key usable in Claude Code, Codex, Cursor and OpenCode
- MiniMax Code agent on web, desktop, CLI, iOS and Android
- H3 video, speech-2.8 TTS, voice cloning, ASR and image models
- Server-side web_search tool and image understanding via MCP
Pros and cons
Pros
- Very low token prices for agentic coding workloads
- Drop-in base URL for tools built on the Anthropic SDK
- One platform covers text, video, speech and image generation
- Flat subscription option with yearly billing at 10 months' price
- Large 1M-token context for big codebases and long sessions
Cons
- M Plan is for individual interactive use, not production traffic
- Usage resets in 5-hour and weekly windows that can cut sessions short
- M2.5 and M2.7 retire on November 18, 2026, forcing migrations
- M3.1 Flash weights are promised but not yet released
- Benchmark claims are vendor-reported, not independently verified
MiniMax pricing
MiniMax uses a freemium model.
| Plan | Price | Includes |
|---|---|---|
| MiniMax Code Free | $0 | Limited M3.1 Flash, image and speech use in MiniMax Code |
| M3.1 Flash API | $0.15 / 1M input tokens | Output $0.60/1M, cached input $0.015/1M; pay-as-you-go |
| M3 API | $0.30 / 1M input tokens | Up to 512k input, after permanent 50% off; output $1.20/1M |
| M Plan Go | $22 / mo | M3.1 Flash, image and speech models; $220 billed yearly |
| M Plan Explore | $55 / mo | 3x Go usage, adds H3 video; $550 billed yearly |
| M Plan Build | $132 / mo | 7.5x Go usage, peak-hour priority; $1,320 billed yearly |
Source: platform.minimax.io · checked on Oct 27, 2026
Are you the maker of MiniMax?
Show a price-verified badge on your site. It displays when we last checked your prices on your own pricing page and links back to this page.
Embed code
<a href="https://happyvalley.dev/tools/minimax"><img src="https://happyvalley.dev/api/badge/minimax?v=price" alt="Prices verified on Happy Valley" height="32"></a>
How MiniMax compares in AI assistants
- Rank by votes
- #1 / 53
- Starting price
- $22
- Free alternatives
- 4
0 votes from members
10% above the category median ($20/mo)
Free or open-source tools in AI assistants
MiniMax ranks #1 of 53 ai assistants tools by community votes and is priced in line with the category.
Common use cases
- 1Running Claude Code or Cursor on a cheaper model through an M Plan key
- 2Powering high-volume coding or browser agents at low per-token cost
- 3Processing long documents or repositories in one 1M-token context
- 4Adding text-to-speech and voice cloning to an app via API
- 5Generating short product or social videos with the H3 models
Best for
Developers and cost-sensitive teams that run agentic coding and automation workloads and want low token prices behind Anthropic- or OpenAI-style APIs.
Not ideal for
Teams that need top-tier frontier models with mature enterprise admin and compliance controls (Claude, Gemini).
Integrations
- Claude Code
- Codex
- Cursor
- OpenCode
- Anthropic SDK
- OpenAI SDK
- AI SDK
- SGLang
Our take
MiniMax is a strong budget choice for agent and coding workloads, and M3.1 Flash pricing undercuts most rivals. Use pay-as-you-go keys for production, treat M Plan as a personal coding subscription, and run your own evaluations before trusting the vendor benchmarks.
FAQ
- What is MiniMax?
- MiniMax is an AI assistants tool: Low-cost agentic coding models, video and speech via API and M Plan.
- How much does MiniMax cost?
- MiniMax uses a Freemium model; paid plans start at $22 per month.
- What are the best alternatives to MiniMax?
- Community-listed alternatives to MiniMax include Claude, DeepSeek, Cursor, Exa.
- What does MiniMax integrate with?
- MiniMax integrates with Claude Code, Codex, Cursor, OpenCode, Anthropic SDK, OpenAI SDK and more.
- Which platforms does MiniMax support?
- MiniMax is available on web, macos, windows, linux, ios, android.
- Who is MiniMax for?
- Developers and cost-sensitive teams that run agentic coding and automation workloads and want low token prices behind Anthropic- or OpenAI-style APIs.
Alternatives to MiniMax
See all 10Claude
▲ 0Anthropic's AI assistant for writing, analysis and code
Freemium · $20Developerschecked Oct 20
DeepSeek
▲ 0Free AI chat app and low-cost API built on open-weight models
FreemiumDeveloperschecked Oct 20
Cursor
▲ 0AI-first code editor with agents, now owned by SpaceX
Freemium · $8Developerschecked Oct 20
Exa
▲ 0Web search and content API built for AI agents and LLM apps
FreemiumDeveloperschecked Oct 22
Deep research reports with every citation checked against source
Open sourceDeveloperschecked Nov 3
Living Brain
▲ 0Shared memory layer for AI agents via API, SDK, CLI and MCP
PaidDeveloperschecked Nov 6
More AI assistants tools
Orphex MCP
▲ 0MCP server that lets AI assistants read and act on ad accounts
Freemium · $40Small businesseschecked Nov 5

Nolyn
▲ 0AI research for ecommerce products, competitors and customer reviews
Freemium · $19Small businessesCharacter.AI
▲ 0Conversational characters and role-play
Freemium · $10Creatorschecked Oct 20
ShogunAI
▲ 0Mac assistant that remembers your work and drafts in your voice
Paid · $62Freelancerschecked Oct 23
Scape
▲ 0AI Mac email client for Gmail that drafts replies from meetings
Freemium · $20Small businesseschecked Oct 30
sideOS
▲ 0Voice-driven Mac agent in the notch with local dictation and meetings
PaidFreelancerschecked Oct 25
Reviews
No reviews yet. Be the first to share real experience.
Sign in to write a review.