Chinese AI API Pricing Landscape — July 2026

Chinese model vendors have spent the last quarter undercutting each other, and the result is the cheapest frontier-adjacent API market in the world. This analysis currently covers 60 models live on our gateway: 25 LLMs and 35 image/video models. Every price below is generated from our public pricing dataset rather than copied into this article by hand. Dataset revision: 2026-07-19; pricing fingerprint: 246472304796c694.

One honest caveat up front: price is not quality. This article tells you what things cost; our benchmark series (task completion rates per model, measured through this same gateway) is the companion that tells you what they're worth.

The sub-$0.20 input tier

5 models charge less than $0.20 per million input tokens. The important comparison inside this tier is output cost: two models with the same cheap input rate can diverge sharply once they start generating long answers.

Two things stand out. First, deepseek-v4-flash and Xiaomi's mimo-v2.5 currently have the same published input and output rates. Second, watch output prices, not just input: qwen3.6-flash sits in the same low-input tier but its output rate is $1, versus $0.2778 for deepseek-v4-flash. That gap dominates generation-heavy workloads.

The 1M-context club

11 of the 25 LLMs in the current catalog publish a 1M-token context window: qwen3.6-flash, qwen3.7-max, qwen3.7-plus, deepseek-v4-flash, deepseek-v4-pro, LongCat-2.0, MiniMax-M3, kimi-k3, mimo-v2.5, mimo-v2.5-pro, glm-5.2. A year ago 1M context was a premium feature; in the Chinese market it is becoming table stakes.

The cheapest published input rate among those million-context models is represented by deepseek-v4-flash at $0.1389 per million input tokens. The most interesting long-context entry is Meituan's LongCat-2.0: an open MoE reasoning model at $0.2778 input / $1.1111 output with the full 1M window — flagship-class specs at flash-tier pricing.

Flagship pricing by vendor

Taking each vendor's highest published output price in the current text catalog as a consistent proxy—not a quality ranking—produces this comparison:

Moonshot's high-speed coding line currently carries the highest output rate in this table at $7.5 per million tokens. Tencent's hy3 is at the other end at $0.5556. That spread is a pricing claim by the vendors, not evidence of the same gap in quality; only task-completion benchmarks can answer whether the premium holds.

Tencent's hy3 deserves a flag: a major-vendor model at $0.1389 input / $0.5556 output per million tokens. Access friction outside China is exactly the gap this gateway exists to close.

Video: per-second vs per-clip, with real math

Video pricing comes in two shapes, and comparing them requires arithmetic vendors don't do for you:

For short drafts, per-generation MiniMax-Hailuo-2.3-Fast is $0.1875 in the representative catalog configuration. Per-second billing (HappyHorse, Wan) can win when you need very short clips or precise duration control. ByteDance's Seedance line has configuration-specific pricing, so the public dataset intentionally marks rows without one representative unit price and points to live pricing instead.

What to pick, by workload

Our stance, from pricing and specs alone (quality benchmarks follow separately):

• Bulk extraction, classification, RAG pre-processing: deepseek-v4-flash or mimo-v2.5 — both currently publish $0.1389/M input with a 1M window.
• Long-document agents on a budget: LongCat-2.0 — 1M context and reasoning-tuned at $0.2778 input / $1.1111 output.
• Multimodal agent loops (image and video input): MiniMax-M3 — per the vendors' own published specs it is currently the only 1M-context natively multimodal model in this catalog.
• Low-cost video drafts among representative per-generation rows: MiniMax-Hailuo-2.3-Fast, currently $0.1875 per generation.
• Premium coding: Moonshot's kimi-k2.7-code line is the price outlier. Benchmark it on your own repository before committing spend.

What this article doesn't tell you

• Quality. A cheap model that fails your task twice can cost more than a premium model that succeeds once. Our task-completion benchmark is the companion piece.
• Stability under load. Chinese vendor APIs vary in rate limits and peak-hour behavior; we route around some of this at the gateway layer, but haven't published numbers yet.
• Tomorrow's prices. Vendors reprice aggressively (usually downward, occasionally promo-then-revert). We now snapshot every price change, so future editions of this article will show trend lines, not just a snapshot.

All 60 models run behind one OpenAI-compatible endpoint with a single key. Download the JSON or CSV, read the methodology, and use the $2 free credit to verify a published price with a real request.