TL;DR Decision Tree:

Want Kimi k2.5 for $0? → OpenCode Zen (ongoing free tier)
Want official IDE for $1? → Kimi Code + AI agent haggling ($0.99-$11.99)
Need agent swarm (beta)? → Allegretto $39/month (Research Preview)
Just testing? → 7-day free trial (card required, auto-renews to $19)
**Critical Update (Feb 3, 2026):** Kilo Code removed free Kimi k2.5 access. OpenCode Zen is now the only ongoing free option. Free tier availability is shrinking—secure access now.

For context on why this pricing exists, see the strategic analysis. For technical comparisons with alternatives, see the three-way comparison.


The Free Path: OpenCode Zen (Ongoing)

Status: Active as of February 3, 2026 Cost: $0 (no credit card required) Best for: Daily coding, testing Kimi capabilities, sustained free access

How to Claim

  1. Visit opencode.ai or install: npm install -g opencode
  2. Select “Zen” hosted option during setup
  3. Create account (no credit card)
  4. Choose Kimi k2.5 from model dropdown

What You Get

  • Kimi k2.5: 76.8% SWE-bench, 256K context, native vision
  • GLM 4.7: Alternative 73.8% SWE-bench model
  • Standard IDE features: code completion, chat, file operations

The Limitations

  • Rate limits: Pay-as-you-go spending caps (exact numbers unpublished)
  • No agent swarm: Base model only, no Allegretto-tier parallel agents
  • Data handling: OpenCode’s hosted tier data policies apply
  • Usage visibility: Spending-based limits, not request counts

When to Upgrade

Free tier supports light-to-moderate usage. Upgrade when:

  • Hitting spending caps daily
  • Needing extended context or faster response times
  • Wanting official Kimi IDE features
  • Requiring agent swarm capabilities (beta)
**Scarcity Alert**: With Kilo Code removing free Kimi k2.5, OpenCode Zen is the last standing ongoing free option. OpenCode may adjust terms as usage concentrates here.

Deep dive: /tools/opencode/ — Self-hosted vs Zen comparison


The Haggling Path: Kimi Code via Kimmmmy ($0.99-$11.99)

Status: Active as of February 3, 2026
Cost: $0.99 to $11.99 (first month, AI-determined) → $19/month ongoing
Best for: Users wanting official IDE at minimum entry cost

The Kimmmmy Pricing Mechanic

Moonshot AI employs an AI agent named “Kimmmmy” to personalize first-month pricing through conversation. This is dynamic pricing, not a fixed promotion.

Observed Range (verified Feb 3, 2026):

  • Lowest observed: $0.99
  • Highest observed: $11.99
  • Typical range: $2.99-$7.99

The Interaction Flow

  1. Visit kimi.com/code
  2. Click “Get Started” or “Bargain” promotion
  3. Interact with Kimmmmy AI agent (multiple conversation rounds)
  4. Receive personalized discount offer
  5. Screenshot the final price before accepting
  6. Complete signup with payment method
  7. Set calendar reminder for day 27 (auto-renews to $19 Moderato)
**Terms Alert**: Per [Kimi's promotion terms](https://www.kimi.com/user/agreement/en/black-friday), "AI-generated content is for reference only." Screenshot your final offer—the system-generated checkout price is what matters, not the conversation.

What Affects Pricing

Based on observational documentation (not prescriptive tactics):

  • Engagement level: More conversation rounds correlate with lower observed prices
  • Interaction quality: Users reporting genuine engagement (questions, not just clicking) report better offers
  • Time/region factors: Pricing may vary by geography and promotional period
**Not Exploitation**: This section documents a product feature, not bugs or exploits. Moonshot designed dynamic AI pricing intentionally. Results vary by user, time, and region.

Post-Trial Reality

  • Auto-renewal: Converts to $19 Moderato after 30 days unless cancelled
  • Cancellation: Must occur before renewal date—no grace period per terms
  • Model lock-in: Kimi k2.5 only, no switching to Claude/GPT

Deep dive: /tools/kimi-code/ — Full trial terms and capabilities


The Trial Path: 7-Day Free (Card Required)

Status: Ongoing
Cost: $0 for 7 days → $19/month auto-renewal
Best for: Evaluating official IDE before financial commitment

How to Claim

  1. Visit kimi.com/code
  2. Select “Moderato” plan
  3. Choose “7-Day Free Trial”
  4. Enter payment details (card required upfront)
  5. Set calendar reminder for day 6 to evaluate cancellation

The Trade-offs

Pros:

  • Full Moderato features during trial
  • No usage limits published for trial period
  • Native IDE experience

Cons:

  • Card required (vs $0 options)
  • Auto-renewal trap (must actively cancel)
  • Same model lock-in as paid tier

When This Makes Sense

Choose 7-day trial over free tiers when:

  • You need to evaluate official Kimi IDE specifically
  • Planning to subscribe if satisfied (card required anyway)
  • Free tier rate limits are already too restrictive
  • Want to test Allegretto features (requires upgrade from trial)

Moderato ($19/month): Daily Driver Tier

What you get:

  • Extended agent quota (2 concurrent tasks)
  • K2.5 with higher speed and extended quota
  • 2048 Kimi Code requests per week
  • 20 deep research uses per month
  • 20 OK Computer uses per month
  • Kimi Code IDE access
  • Slides visual mode with Nano Banana

What you don’t get:

  • Research Preview Agent Swarm (Allegretto+)
  • API access (subscription ≠ API credits)
  • Model switching (Kimi k2.5 only)

Best for: Individual developers with moderate daily usage who want the official Kimi experience without API complexity.

Break-even vs API:

  • Moderato unlimited vs API at 10M+ tokens/month
  • Subscription wins on convenience, not pure cost for light usage

Allegretto ($39/month): Agent Swarm Beta

**First Impressions Pending**: We have not tested Allegretto's Research Preview Agent Swarm. This documents Moonshot's claims only.

What you get (beyond Moderato):

  • 2x agent quota
  • 2x usage quota for K2.5
  • Kimi Code 3.5x quota
  • Research Preview Agent Swarm (up to 100 sub-agents, beta)
  • Unlimited Slides visual mode
  • Refund for Moderato→Allegretto upgrade prorated

Agent swarm capabilities (per Moonshot):

  • “Multiple agents, multiplied productivity”
  • Use cases: massive search, long writing, batch tasks
  • Parallel execution with hierarchical coordination

Best for: Power users needing parallel agent execution, content creators doing batch workflows, teams testing swarm architectures.

The bet: You’re paying $20/month extra for beta features that may not be fully mature. Evaluate Moderato first, upgrade if you hit agent limits.

Vivace ($199/month): Team/Heavy Usage

What you get:

  • 10x agent quota
  • 2x agent multi-tasking
  • Priority access during peak hours
  • 10x usage quota
  • Kimi Code 17.5x quota
  • Research Preview Agent Swarm
  • Team collaboration features

Best for: Engineering teams, agencies, heavy production workflows.


OK Computer Rewards: Referral Mechanics

The OK Computer system provides usage credits for referrals and participation.

How It Works

For Referrers (Subscribed Users):

  • Generate exclusive discount link via Kimmmmy interaction
  • Share with non-subscribed users
  • Receive 10 OK Computer units when invitee subscribes
  • One-time reward per referrer (maximum 10 units total)

For All Users (More Perks Program):

  • Share invitation link from “More Perks” page
  • 1 OK Computer unit when friend participates in event
  • 5 OK Computer units when friend completes subscription
  • Cumulative rewards based on real-time event page data

Critical Limitations

**Expiration Policy**: OK Computer credits expire aggressively: - Credits earned before Jan 1, 2026: Expired Jan 31, 2026 - Credits earned on/after Jan 1, 2026: Valid only for calendar month granted, expire at month start

Usage: OK Computer units apply to “OK Computer” feature only—not general Kimi k2.5 usage.

Maximizing Value

  1. Time referrals strategically: Credits expire monthly, so refer near month-start
  2. Verify receipt: Check free OK Computer balance in-app; email notifications may not arrive
  3. Don’t hoard: Use credits immediately—expiration is automatic and non-negotiable
  4. Track via event page: Real-time cumulative status shown on promotion page

Geographic Restrictions

Not available in:

  • Mainland China (explicitly excluded per terms)
  • Some regions restricted due to “payment capabilities or policy regulations”

Verification required: Check signup flow for your region before relying on any specific pricing.

Data handling: Moonshot AI is a Chinese company. Review data handling risks before uploading sensitive code or data.


Decision Framework: Which Path for You?

Your SituationRecommended PathCostCommitment
Testing/experimentingOpenCode Zen$0None
Want official IDE cheapKimi Code + haggle$0.99-$11.9930 days
Need to evaluate officially7-day trial$07 days
Daily driver, no swarm needsModerato$19/moOngoing
Power user, batch workflowsAllegretto$39/moOngoing
Team/enterpriseVivace$199/moOngoing

The Risk-Averse Strategy

  1. Start with OpenCode Zen (free, immediate, no CC)
  2. If you hit limits, try Kimi Code haggling ($0.99-$11.99, 30-day commitment)
  3. If satisfied, continue Moderato ($19/mo) or upgrade to Allegretto ($39/mo) if you need swarm

Never pay $19+ without testing free options first.


Quick Command Reference

OpenCode Zen:

1
2
3
npm install -g opencode
opencode --zen
# Select Kimi k2.5 during setup

Kimi Code (if subscribed):

1
2
curl -L code.kimi.com/install.sh | bash
kimi --version

Check subscription status:

  • Web: kimi.com/code → Account → Subscription
  • CLI: kimi account (if available)

Value Section:

Analysis:

Model & Tool Guides:

Verification & Risks:


Last verified: February 3, 2026
Sources: kimi.com/code pricing page (screenshot verified), Kimi Promotion Terms (black-friday agreement), OpenCode documentation, Kilo Code dashboard (free tier removal confirmed)
Evidence level: High (primary source screenshots + official documentation)