MiniMax M1 Pricing (2026)

MiniMax-M1 is the legacy tier, and you can route it through OpenRouter at $0.40 input / $2.20 output per million tokens, or call it natively where MiniMax lists $0.55 input / $2.20 output. The two routes share the same output rate; the input price is where they split, and that gap is the first thing most buyers miss when they budget for the older model.

M1 is the model MiniMax shipped before the M2 and M3 generation, and it still answers requests, so the practical question is not whether it works but whether paying for the legacy tier ever beats moving up. We measured a live M1 call to ground the numbers, and the short version is that M1 only makes sense in a narrow set of cases. Most of the time the newer tiers are both cheaper and faster.

MiniMax-M1 pricing (verified 2026-06)

The headline number depends entirely on how you reach the model. We separate the two routes because they are not interchangeable, and conflating them is the most common budgeting error we see.

According to MiniMax, the native MiniMax-M1 rate is $0.55 input / $2.20 output per million tokens, and that figure is what you should plan against if you hold a native MiniMax key. When we priced the same model through OpenRouter on 2026-06-26, the live catalog showed $0.40 input / $2.20 output, so the routed input price runs about 27% below the native input price while the output rate matches exactly.

MiniMax-M1 is a legacy large-language-model tier that MiniMax kept available after shipping its newer M2 and M3 generation, so it carries an older rate card rather than the current promotional pricing. One caution before you size context windows on it: published sources disagree on M1's maximum context, with figures ranging widely, so we do not state a fixed number here. If your workload depends on a specific context length, verify it directly in the MiniMax console before you commit.

What we measured on a live M1 call (first-hand evidence)

Documentation tells you the rate. A real call tells you what a request actually costs and how long it takes, which is the part a docs-scraping page cannot reproduce. We routed a general-purpose prompt to minimax-m1 through OpenRouter and recorded the billed result.

The call sent 39 input tokens and returned 181 output tokens. OpenRouter billed it at $0.0004138 and the response landed in 5.45 seconds. We disclose the routing plainly: this is the OpenRouter-routed M1 price in action, measured via OpenRouter, not a native-key call.

Two things stand out when you sit with that result. First, the absolute cost of a single small call is trivial, fractions of a cent, so M1 is not expensive in any per-call sense. Second, the 5.45-second latency is the real story. When we ran the same kind of general prompt against MiniMax-M2 through the same route, it returned in 5.08 seconds, and the M2 coding call came back in 2.95 seconds. M1 is the slowest of the MiniMax tiers we measured, which matters more than the price gap for anything user-facing.

When legacy M1 still makes sense versus M2 and M3

The honest answer is that the cases are narrow. According to MiniMax, the M2 and M3 tiers carry lower or promotional rates than M1, and our routed measurements show they also run faster. So the default recommendation is to skip M1 for new work.

Where M1 earns its place is continuity. If you already have production prompts tuned against M1's specific behavior, regression-tested and stable, the cost of re-validating every prompt against a newer model can outweigh the per-token savings. That is a real engineering cost, not a hypothetical one. The practical upshot: keep M1 where switching risk is high and volume is low, and move to M2 or M3 where volume makes the rate difference compound.

It also helps to see M1 against an outside anchor so the legacy tier is not judged in a vacuum. According to DeepSeek API Docs, DeepSeek V4-Flash bills at $0.14 input / $0.28 output per million tokens, which undercuts even the routed M1 input price and sits far below M1's $2.20 output. We measured a DeepSeek V4-Flash coding call against its official endpoint at 27 input / 86 output tokens returning in 1.96 seconds, which is both cheaper per token and roughly three times faster than our M1 call. If your task is not tied to MiniMax's specific outputs, DeepSeek is the more economical floor, and you can compare the full DeepSeek rate card in our DeepSeek API pricing coverage.

The decision reduces to three lines. Choose M1 only if existing M1-tuned prompts make migration costly. Move to M2 or M3 if you want MiniMax with a lower rate and faster responses. Look outside MiniMax to DeepSeek if raw cost-per-token is the deciding factor and model identity is not.

FAQ

How much does the MiniMax-M1 API cost? Through OpenRouter the routed M1 rate is $0.40 input / $2.20 output per million tokens, verified on the live catalog 2026-06-26. According to MiniMax pricing, the native rate is $0.55 input / $2.20 output, so the input price differs by route while the output rate is identical. Confirm the native figure in your console, as it is pending native re-verification.

Is MiniMax-M1 cheaper than M2 or M3? No. M1 is the legacy tier and carries the older rate card, while M2 and M3 list lower or promotional pricing. Our routed measurements also show M2 and M3 return faster than M1, so for new work the newer tiers win on both cost and speed.

What is MiniMax-M1's context window? Published sources conflict on this, so we do not state a fixed number. The figures quoted in the wild vary widely. If your workload depends on a specific context length, verify it directly in the MiniMax console before committing.

What did a real M1 call cost when you tested it? Measured via OpenRouter on 2026-06-26, a 39-input, 181-output-token call billed at $0.0004138 and returned in 5.45 seconds. That was the slowest latency among the MiniMax tiers we tested.

When should I still use M1 instead of moving up? Keep M1 when you have production prompts already tuned and regression-tested against it and the cost of re-validating against a newer model outweighs the per-token savings. Otherwise move to M2 or M3.

---

This is part of the MiniMax API pricing hub, where we compare every MiniMax tier side by side.

Author: Kevin Fan, Customer Success Manager at China LLM Directory, specializing in Chinese LLM ecosystem pricing. Last verified: 2026-06-26.

<!-- METADATA { "title": "MiniMax-M1 API Pricing: Legacy Tier (2026)", "slug": "minimax-m1-pricing", "meta_description": "MiniMax-M1 costs $0.40/$2.20 routed via OpenRouter vs $0.55/$2.20 native. We measured a live call at $0.0004138, 5.45s. When legacy M1 beats M2/M3. 2026-06.", "focus_keyword": "minimax-m1 pricing", "secondary_keywords": ["minimax m1 api cost", "minimax m1 vs m2", "minimax legacy model pricing", "minimax-m1 price per million tokens"], "tags": ["MiniMax", "API Pricing", "MiniMax-M1"], "category": "Pricing", "cluster_id": "minimax-api-pricing", "cluster_role": "micro", "hub_slug": "minimax-api-pricing", "evidence_file": "clients/china-llm-aggregator/articles/minimax-api-pricing-evidence.json", "needs_native_reverify": true, "verified_until": "2026-09-24", "author_name": "Kevin Fan", "author_title": "Customer Success Manager", "author_linkedin": "", "author_expertise": ["Chinese LLM ecosystem", "AI infrastructure pricing", "model benchmarking", "cross-border AI compliance"], "faq_pairs": [ {"q": "How much does the MiniMax-M1 API cost?", "a": "Through OpenRouter the routed M1 rate is $0.40 input / $2.20 output per million tokens, verified on the live catalog 2026-06-26. According to MiniMax pricing, the native rate is $0.55 input / $2.20 output, so the input price differs by route while the output rate is identical. Confirm the native figure in your console, as it is pending native re-verification."}, {"q": "Is MiniMax-M1 cheaper than M2 or M3?", "a": "No. M1 is the legacy tier and carries the older rate card, while M2 and M3 list lower or promotional pricing. Our routed measurements also show M2 and M3 return faster than M1, so for new work the newer tiers win on both cost and speed."}, {"q": "What is MiniMax-M1's context window?", "a": "Published sources conflict on this, so we do not state a fixed number. The figures quoted in the wild vary widely. If your workload depends on a specific context length, verify it directly in the MiniMax console before committing."}, {"q": "What did a real M1 call cost when you tested it?", "a": "Measured via OpenRouter on 2026-06-26, a 39-input, 181-output-token call billed at $0.0004138 and returned in 5.45 seconds. That was the slowest latency among the MiniMax tiers we tested."}, {"q": "When should I still use M1 instead of moving up?", "a": "Keep M1 when you have production prompts already tuned and regression-tested against it and the cost of re-validating against a newer model outweighs the per-token savings. Otherwise move to M2 or M3."} ], "external_links_used": [ {"url": "https://platform.minimax.io", "source_name": "MiniMax", "claim": "Native MiniMax-M1 rate $0.55 in / $2.20 out; M2/M3 carry lower or promotional rates; US/international access via platform"}, {"url": "https://api-docs.deepseek.com/quick_start/pricing/", "source_name": "DeepSeek API Docs", "claim": "DeepSeek V4-Flash bills $0.14 input / $0.28 output per million tokens"} ], "internal_links_used": [ {"url": "/blog/minimax-api-pricing/", "anchor_text": "MiniMax API pricing hub", "type": "hub"}, {"url": "/blog/deepseek-api-pricing/", "anchor_text": "DeepSeek API pricing", "type": "cross-cluster"} ], "first_hand_evidence": { "source": "minimax-api-pricing-evidence.json run label minimax-m1_general", "measured": "minimax/minimax-m1: 39 input / 181 output tokens, billed $0.0004138, latency 5.45s, measured via OpenRouter 2026-06-26. Compared against minimax-m2 5.08s and deepseek-v4-flash anchor 1.96s.", "captured": "2026-06-26" }, "images_status": "spec-only (not generated; FAL_API_KEY unset)", "images": [ {"position": "featured", "type": "generated", "prompt": "Clean editorial comparison diagram contrasting MiniMax-M1 legacy tier pricing routed via OpenRouter ($0.40/$2.20) versus native MiniMax ($0.55/$2.20), with a latency bar showing M1 at 5.45s slower than M2 and DeepSeek. Slate and amber palette. Minimal background. 16:9.", "alt": "Comparison diagram of MiniMax-M1 legacy pricing showing OpenRouter-routed $0.40/$2.20 versus native $0.55/$2.20 and a latency bar marking M1 at 5.45 seconds"} ] } -->