Global AI Token Comparison
Global AI Token Plan comparison home

Xunfei Token Plan Pricing, Top-up, Deals & Official Entry

Xunfei Astron Token Plan pricing splits into two lines: Astron Coding Plan personal tiers (High-efficiency ¥199/mo, Speed ¥999/mo; Worry-free and Pro discontinued) bill by request caps, while Token Plan Team seats (Standard ¥160, Advanced ¥420, Premium ¥1,200 per seat/mo) bill by Credits plus TPM. This page gathers official access, top-up, deals, API keys / Base URLs, the Spark X2, GLM, Kimi, MiniMax, DeepSeek and Qwen model pools, request quotas and Credits/TPM rules, plus Claude Code, Cursor, OpenCode, OpenClaw and TRAE setup, to judge whether personal AI coding or team project purchasing beats the alternatives.

Token Plan19 models · personal + teamAstron Coding Plan
SubscriptionToken API

Xunfei Token Plan Latest Updates

Pricing

Xunfei Astron Token Plan Retires Worry-free & Pro: ¥199 High-efficiency vs ¥999 Speed Pricing Comparison

Xunfei Astron Token Plan has retired the ¥19/mo Worry-free and ¥39/mo Pro tiers; only ¥199/mo High-efficiency and ¥999/mo Speed (first month ¥699) remain. This pricing comparison breaks down the two tiers' quotas, the July 23 model deduction coefficients, whether upgrading from an old plan is worth it, and how to connect Claude Code, Cursor and TRAE.

View all updates

Below is the complete pricing comparison for Xunfei Token Plan, covering 7 plan tiers, core model capabilities, quotas, and official entry. All data is sourced from the official website to help you decide whether it’s worth it.

Last updated：

Xunfei Token Plan Price Comparison: Core Models

Spark X2 AgentSpark X2Spark-X2-FlashAutoGLM-5.2GLM-5.1GLM-5DeepSeek-V4-ProDeepSeek-V4-FlashDeepSeek-V3.2Kimi-K2.7-CodeKimi-K2.6Kimi-K2.5MiniMax-M2.5Qwen3.5-397B-A17BQwen3.6-35B-A3BQwen3.5-35B-A3BQwen3-Coder-Next-FP8GLM-4.7-Flash
Spark X2 Agent

xsparkx2agent—Agent-enhanced Spark X2 for autonomous coding agents; Speed tier only.

Spark X2

xsparkx2—iFlytek deep-reasoning flagship for math and agents; available on High-efficiency tier.

Spark-X2-Flash

xsparkx2flash—value Spark line with stronger agent and coding skills.

Auto

auto—console smart-routing model, auto-selects optimal base model per task; High-efficiency/Speed tiers.

GLM-5.2

xopglm52—Zhipu's latest flagship with significant long-horizon improvements; Speed tier.

Xunfei Token Plan Price Comparison: Plans

无忧版
Discontinued
Price Comparison
¥19/mo
Unlimited requests
Monthly
¥19/mo
Auto quarterly
¥16.6/mo
12% off quarterly
Usage
⚠️ Discontinued, no longer available for new purchase. Worry-free's hallmark was unlimited request count for batch light tasks, learning, and small edits—but only light modelIds like Qwen3.5-35B-A3B, Qwen3-Coder-Next-FP8, GLM-4.7-Flash—not unlimited heavy agents. Now superseded by High-efficiency/Speed.
Models
Each question deducts actual model calls (simple ~3–10, complex 10–30+)—track usage on the MaaS subscription page.
Highlights
Suited to AI coding as support, not a primary workbench—upgrade to High-efficiency or Speed for long complex chains.
Same-tier stacking allowed—each purchase gets its own API key and quota; 31-day validity.
Best for
Students, indie developers, and lightweight batch-task users
专业版
Discontinued
Price Comparison
¥39/mo
~18,000/mo
Monthly
¥39/mo
Auto quarterly
¥37/mo
5% off quarterly
Usage
⚠️ Discontinued, no longer available for new purchase. Pro's original cap was ~1,200/5h, ~9,000/wk, ~18,000/subscription-month—now superseded by High-efficiency/Speed.
Models
The former Pro tier could switch Auto, Spark-X2-Flash, GLM-5, Kimi-K2.5, MiniMax-M2.5, DeepSeek-V3.2 plus light Qwen/GLM modelIds for mainstream coding.
Highlights
Quotas start at payment with 31-day subscription cycle—track usage on the MaaS subscription page.
Upgrade to High-efficiency or Speed with prorated difference—no downgrade; peak throttling possible.
Best for
Medium-to-high-frequency developers using AI coding as a daily tool
高效版
HeavyRecommended
Price Comparison
¥199¥199/mo
~6,000/5h · 45,000/wk · 90,000/mo
Monthly
¥199/mo
Auto quarterly
¥199¥179.3/mo
10% off (everyday)
Usage
High-efficiency at ¥199/mo is ~5× Pro: 6,000/5h, 45,000/wk, 90,000/subscription-month; new purchases from July 23, 2026 use model coefficients (Spark X2 Agent x2, GLM-5.2 x5, Kimi-K2.7-Code x5 etc.)—for heavy multi-project and long agent chains.
Models
Adds GLM-5.1, Spark X2, Kimi-K2.6, DeepSeek-V4, and more atop Pro's pool—for higher engineering and output demands.
Highlights
If AI coding is a formal productivity system—not occasional support—this tier’s advantage over Pro becomes meaningful.
Same 31-day validity—upgrade from lower tiers supported; no downgrade or refund.
Best for
Heavy developers, high-frequency enterprise users, and multi-project teams
速通版
Speed
Price Comparison
¥999¥999/mo
First month ¥699 · ~30,000/mo
Monthly
¥999¥699/mo (first)
First-month promo
Auto quarterly
¥999¥899/mo
10% off (everyday)
Usage
Speed at ¥999/mo (limited first-month ¥699), ~30,000/subscription-month request cap with full model coefficients—for high-efficiency, high-intensity dev; addresses peak queuing and timeout.
Models
Full access to Spark X2 Agent, GLM-5.2, Kimi-K2.7-Code, DeepSeek-V4-Pro, and all modelIds—speed quota ensures faster responses.
Highlights
Same 31-day validity—upgrade from lower tiers supported; no downgrade or refund.
Best for
All-day heavy developers and teams needing priority response and full model coverage
标准成员
Standard seat
Price Comparison
¥200¥160/seat/mo
20,000 credits/mo · 2M TPM
Usage
Team entry tier: ¥160/seat/mo (limited 20% off, regular ¥200) with 20,000 credits/mo and a 2M TPM cap, billed per seat (0–200 seats).
Models
All 17 models available (Spark X2 family, GLM-5/5.1/5.2, DeepSeek-V4/V3.2, Kimi-K2.5/K2.6, MiniMax-M2.5, Qwen3.5/3.6, Qwen3-Coder-Next-FP8, GLM-4.7-Flash); each model deducts credits per its rule.
Highlights
Skill benefits are bundled free and don't consume credits: hyper-realistic TTS, one-shot voice cloning, offline ASR, image/PDF OCR, receipt recognition, official-document proofreading, translation, and image understanding.
Also includes an Astronclaw basic membership (worth ¥168 × 1 month); 31-day validity, non-refundable.
Best for
Small teams with light collaboration
高级成员
Advanced seat
Price Comparison
¥600¥420/seat/mo
60,000 credits/mo · 3M TPM
Usage
Advanced seat: ¥420/seat/mo (limited 30% off, regular ¥600) with 60,000 credits/mo and a 3M TPM cap, billed per seat (0–200 seats).
Models
All 17 models available with per-model credit deduction—for teams needing a higher credit budget and concurrent throughput.
Highlights
Skill benefits bundled free (no credit cost) plus an Astronclaw basic membership (¥168 × 1 month); 31-day validity, non-refundable.
Best for
Teams with mid-to-high concurrency and heavy daily AI use
尊享成员
Premium seat
Price Comparison
¥2000¥1,200/seat/mo
200,000 credits/mo · 5M TPM
Usage
Premium seat: ¥1,200/seat/mo (limited 40% off, regular ¥2,000) with 200,000 credits/mo and a 5M TPM cap, billed per seat (0–200 seats).
Models
All 17 models with the highest credit and TPM caps—for large teams, high-concurrency workloads, and budget-conscious bulk purchases.
Highlights
Skill benefits bundled free (no credit cost) plus Astronclaw basic membership (¥168 × 1 month); 31-day validity, non-refundable. For enterprise bulk purchase contact [email protected].
Best for
Large teams and enterprise bulk purchase

Xunfei Token Plan Price Comparison: Quota & Usage

Product type
Token Plan · Coding Plan
Official product line
Astron Coding Plan
5-hour limit
6,000/5h (High-efficiency)
Weekly limit
45,000/wk (High-efficiency)
Monthly limit
90,000/mo (High-efficiency) · 30,000/mo (Speed)

Xunfei Token Plan Price Comparison: Notes

  • Both Worry-free (¥19/mo) and Pro (¥39/mo) are discontinued; current tiers are High-efficiency ¥199/mo (6,000/5h, 45,000/wk, 90,000/mo) and Speed ¥999/mo (first-month ¥699, ~30,000/mo). Valid 31 days—stacking and upgrades supported, no downgrade/refund.
  • Coding Plan lists 19 modelIds; use astron-code-latest by default and switch the base model on the MaaS subscription page (effective in 1–3 min). New purchases from July 23, 2026 apply model coefficients such as GLM-5.2 / DeepSeek-V4-Pro / Kimi-K2.7-Code x5, Spark X2 / Spark X2 Agent / Auto x2, and Spark-X2-Flash x1.
  • From 2026-06-18 off-peak coefficient: weekdays 08:00–22:00 at 1.0; nights/weekends/holidays at 0.8—off-peak use stretches caps (e.g. 6,000/5h ≈ 7,500 effective).
  • Use astron-code-latest (console model switch in 1–3 min) or any of 19 documented modelIds for live switching in tools.
  • Once Token Plan Team seats are active, users call any of the 17 models with credits deducted per model rules; Skill benefits are bundled free and don't consume credits. See the official TokenPlan doc (xfyun.cn/doc/spark/TokenPlan.html) for credit rules; enterprise bulk purchase via [email protected].

Xunfei Token Plan Price Comparison: Tools & Integration

Claude CodeCursorOpenCodeOpenClawCodex CLITRAE

Xunfei Token Plan Pricing, Top-up, Deals, Quota, Usage, Setup & Updates

View more updates

Page published：

Xunfei Token Plan Pricing, Top-up, Deals & FAQ

AI Token Plan

Global AI Token Comparison

AI Token Plan is not just a collection of vendor links. It places 41+ domestic and international AI platforms into one comparison framework, covering 395+ model entries plus common plan types such as Token Plans, Coding Plans, IDE Tools and LLM APIs.

When you need to compare AI subscription pricing, coding plan quota, API usage cost or official deal entry points, AI Token Plan brings official prices, plan tiers, usage rules, model capabilities and tool integrations into one place, reducing the need to check multiple vendor sites manually.

Data is continuously organized as vendor pricing pages, plan pages and product documentation change, making the homepage a pricing comparison entry point while detail pages explain whether each vendor plan fits individual developers, team purchasing or long-term API usage.

Prices come from each platform's official site and may change at any time; the official price prevails.

© 2026 AI Token Plan · All rights reserved · First published June 18, 2026 · 64 days running · Sitemap