Last updated: 2026-07-17 · Report outdated data on GitHub

MiMo API Pricing — July 2026

Frontier-class reasoning at 5–20% of the cost of closed-source alternatives. Permanently cut ~99% in May 2026. All prices in USD per 1 million tokens.

Bottom line: ~99% off in 14 months

MiMo-V2.5-Pro now sits at $1.00 per 1M input / $3.00 per 1M output — down from roughly $100+ per 1M input at its 2025 launch. The same V2.5-Pro is ~93% cheaper than Claude Opus 4.5 on input tokens, and ~96% cheaper on output. V2-Flash undercuts Claude Haiku 3.5 by ~38% on input and ~63% on output. Across the lineup, MiMo lands in the price band of DeepSeek V4 Pro and Qwen3 Max — two of the cheapest frontier models — but with substantially better long-context performance (1M tokens on V2.5-Pro vs 128k on V4 Pro).

All MiMo weights are MIT-licensed. Self-hosting eliminates per-token cost entirely for teams with GPU infrastructure.

→ Use the Token Cost Calculator (real-time USD/CNY)

1. The V2.5 lineup (full pricing)

All prices per 1 million tokens. List rates as of July 17, 2026.

Model Params Context Best for Input / 1M Output / 1M Change
MiMo-V2.5-Pro 1T+ MoE 1M Agent, long-context, reasoning $1.00 $3.00 ▼ 99% (May 2026)
MiMo-V2-Flash 309B / 15B active 56k Coding agents, fast chat $0.50 $1.50 ▼ 99% (May 2026)
MiMo-V2.5-Omni TBD TBD Multimodal (text / image / video / audio) TBD TBD Coming 2026
MiMo-V2.5-TTS TBD N/A Bilingual speech synthesis (CN/EN) Commercial license Coming 2026
MiMo-7B 7B dense 32k Edge / on-device (phones, cars, IoT) $0.00 $0.00 MIT weights, free
MiMo Code Agent layer Infinite Terminal-native coding agent Free app + API usage Open source

2. Side-by-side: same workload, 5 leading APIs

All prices list (USD per 1M tokens). Volume discounts, prompt caching, and batch API rates not included. Verified July 17, 2026 against provider list pages.

Provider Model Input / 1M Output / 1M vs MiMo-V2.5-Pro (input)
Xiaomi MiMo-V2.5-Pro $1.00 $3.00 — baseline —
Xiaomi MiMo-V2-Flash $0.50 $1.50 50% cheaper input
DeepSeek DeepSeek V4 Pro $0.435 $0.87 ~57% cheaper input
DeepSeek DeepSeek-R1 $0.55 $2.19 ~45% cheaper input
Alibaba Qwen3 Max $0.60 $2.40 ~40% cheaper input
OpenAI GPT-4o mini $0.15 $0.60 ~85% cheaper input (smaller model)
OpenAI GPT-5 $1.25 $10.00 ~25% more expensive input
Anthropic Claude Haiku 3.5 $0.80 $4.00 ~20% cheaper input (smaller model)
Anthropic Claude Opus 4.5 $15.00 $75.00 15× more expensive input

Reading the table: MiMo-V2.5-Pro matches Claude Opus 4.5 on long-context reasoning, beats Claude Haiku 3.5 on coding (73.4% vs 65% on SWE-Bench Verified), and is priced between DeepSeek V4 Pro and GPT-5 on input. The 1M context window (vs 128k for V4 Pro) is the deal-breaker for legal / scientific / code-review workloads.

3. Pricing history (timeline)

May 2026

Permanent 99% price cut. V2.5-Pro drops to $1/$3 per 1M, V2-Flash to $0.50/$1.50. Xiaomi cites "ecosystem expansion over short-term margin" as the rationale. This is the largest single price cut in the frontier-model market in 2026.

March 2026

V2.5-Pro public beta. Initially priced at $5/$15 per 1M during a 60-day beta. Reserved capacity for 5,000 enterprise accounts.

November 2025

V2-Flash GA. Launched at $0.50/$1.50 per 1M — already the cheapest MoE model at its parameter class.

June 2025

MiMo-7B open-weights release. MIT licensed. Free to use, modify, and self-host. Zero per-token cost on any hardware that fits (INT4 = 3.5 GB VRAM).

4. Where to buy the MiMo API

5. Pricing FAQ

Why did MiMo cut prices 99% in May 2026?

Xiaomi's stated rationale: V2.5-Pro's throughput at 1000+ tokens/second (UltraSpeed mode, FP8 + speculative decoding) reduced marginal inference cost enough that the $1/$3 price still produces healthy margin. The cut is positioned as a market-share play — Xiaomi wants to be the default model for cost-sensitive agent workloads.

Are there volume discounts?

Not at the official list rate. Enterprise contracts (1B+ tokens/month) are negotiated directly with Xiaomi's sales team — typically 20–40% off list. OpenRouter also offers pay-as-you-go without volume tiers.

What about prompt caching and batch API?

MiMo's public API as of July 2026 does not yet expose prompt caching or batch API rates. If you need either, OpenRouter adds prompt caching on top with a 25% cache-write discount / 90% cache-read discount on MiMo variants.

Is self-hosting cheaper than the API?

Depends on volume. Break-even for MiMo-7B (the smallest self-hostable model) is around 50M tokens/month on a single A10G. For V2-Flash and V2.5-Pro, you'd need multi-GPU infrastructure — break-even is 500M+ tokens/month. See deployment guide →

Is V2.5-Omni and V2.5-TTS pricing finalized?

Not yet — both models are listed as TBD on the public API as of July 2026. Weights are MIT-licensed (Omni) or commercially licensed separately (TTS). Sign up for platform.xiaomimimo.com to get notified when API tiers launch.

Related pages

Unofficial community resource. Not affiliated with Xiaomi Inc. Prices last verified July 17, 2026 against provider list pages. For mission-critical budgeting, always confirm with the provider's pricing page.