Kimi Code is Moonshot’s current coding membership path, and its current model configuration lists Kimi K3 and Kimi K2.7 Code. It supports the official CLI, selected IDE workflows, and third-party coding agents through tool-facing model IDs.

Use Kimi K3 for the newest flagship model evidence and Kimi K2.7 Code for cheaper pay-as-you-go coding economics. Kimi Code membership is a separate quota and credit product.

Current Status

ItemJuly 18 read
Product models shownKimi K3 and Kimi K2.7 Code
CLICurrent Node.js-based Kimi Code CLI
Tool-facing model IDsk3, kimi-for-coding, kimi-for-coding-highspeed
ProtocolsOpenAI-compatible and Anthropic-compatible endpoints
Membership accountingShared Kimi credits plus Kimi Code quota/rate windows
API platform billingSeparate pay-as-you-go Kimi Platform product

The tool-facing IDs are not the same contract as explicit Kimi Platform API IDs such as kimi-k3 or kimi-k2.7-code. Record the displayed backend model when reproducibility matters.

Install and Authenticate

1
2
curl -fsSL https://code.kimi.com/kimi-code/install.sh | bash
kimi

Use /login for the official OAuth flow. Membership users can also create API keys for supported third-party tools.

Current Kimi docs list membership/tool routes separately from the Kimi Platform API. For explicit API model identity, use kimi-k3 or kimi-k2.7-code through the Kimi API platform.

  • OpenAI-compatible base URL: https://api.kimi.com/coding/v1
  • Anthropic-compatible base URL: https://api.kimi.com/coding/
  • Model IDs: k3, kimi-for-coding, or kimi-for-coding-highspeed depending on client and entitlement

Keep the tool’s real identity/User-Agent. Kimi warns that spoofing the client identity can suspend membership benefits.

CLI, IDE, and Third-Party Tools

Kimi documents code reading, file editing, command execution, web search, goals, MCP, skills, plugins, hooks, and subagents in the current CLI.

The VS Code path needs a caveat: current docs say new VS Code extension installations are limited to legacy Python CLI users while the TypeScript CLI integration is adapted. Existing users can continue, and other editors can connect through the CLI’s ACP protocol.

Third-party integrations include Claude Code, Roo Code, OpenCode, OpenClaw, and compatible agent frameworks. Compatibility does not mean unlimited use or a direct K2.7 pay-as-you-go entitlement.

Quotas and Credits

Two layers apply:

  1. Kimi Code quota refreshes every seven days from the subscription date; unused weekly quota does not carry over.
  2. A rolling five-hour frequency window can rate-limit bursts even when weekly quota remains.

All logged-in devices and Kimi Code API keys share the same account quota. Kimi also says membership features now use a unified monthly credit pool, including Kimi Code, and that exhausting monthly credits can freeze new Code work until refresh or upgrade.

Do not preserve older fixed request-count promises as current guarantees. Check the console and subscription page for the account’s remaining credits, frequency status, and next refresh.

Membership vs Pay-As-You-Go API

QuestionKimi Code membershipKimi Platform API
BillingSubscription credits and quotasPay as you go
Model IDk3, kimi-for-coding, or kimi-for-coding-highspeedkimi-k3, kimi-k2.7-code, or kimi-k2.7-code-highspeed
Best fitTerminal/IDE agent programmingProduct integration and measured API workloads
ReproducibilityBackend alias can upgradeExplicit model ID
LimitsWeekly, five-hour, and monthly credit controlsPublished API rates and provider rate limits
K3 reasoninglow, high, or max through membership model ID k3max only in the current pay-as-you-go API guide
K3 contextModerato: up to 256K; Allegretto and above: up to 1M1M model context, subject to API request limits

Current K3 API pricing is $0.30 cache-hit input, $3.00 cache-miss input, and $15.00 output per million tokens. Current K2.7 Code API pricing is $0.19 cache-hit input, $0.95 cache-miss input, and $4.00 output per million tokens. HighSpeed doubles K2.7 token prices. Do not use API rates to calculate the value of a membership quota without measured consumption.

Do not transfer Kimi Code settings to the API by analogy. Membership k3 currently supports low/high/max reasoning and plan-dependent 256K/1M context; pay-as-you-go kimi-k3 currently uses always-on, max-only reasoning. Start a new session when changing the membership model or reasoning effort so the old context cache is not mistaken for a reusable K3 session.

Model Evidence

benchmark artifact

Kimi Code and Current Alternatives

ModelProviderStatusContextInput priceOutput priceCoding signalTool-use signalBenchmark evidenceSpeedVerdictSourcesChecked
Kimi K3Moonshot AIactive
Active Kimi API/product flagship; official weights and serving materials released under the Kimi K3 License.
1M$3.00 / 1M$15.00 / 1MMoonshot reports strong max-reasoning launch-suite coding and agent results; AIHackers repo eval is not verified.Kimi API and Kimi Code support K3; Kimi docs warn to preserve full assistant history and avoid mid-session model switching.
  • Artificial Analysis Intelligence Index v4.1: 57 (independent)
  • Artificial Analysis output speed: 62 tokens/s (independent)
  • Moonshot launch benchmark suite: vendor-reported max-reasoning table (vendor)
  • AIHackers repo eval: not verified (site-owned)
Artificial Analysis reports 62 output tokens/s and flags high verbosity; measure total output cost per accepted task.Test as Kimi's newest 1M-context frontier-adjacent lane; keep K2.7 Code for cheaper routine Kimi coding until K3 passes local CAR tests.Kimi K3 launch blog [archive], Kimi K3 quickstart [archive], Kimi K3 API pricing [archive], Kimi K3 weights and license [archive], Kimi current model list [archive], Kimi Code model configuration, Artificial Analysis: Kimi K3 [archive]2026-08-01
Kimi K2.7 CodeMoonshot AIactive
Cheaper routine Kimi coding API lane; HighSpeed is the same model at higher token prices.
256K$0.95 / 1M$4.00 / 1MKimi K2.7 Code remains the lower-cost 256K coding lane after K3; independent normalized benchmarks are not imported.OpenAI-compatible API; thinking mode required in the documented K2.7 Code quickstart.
  • Program-Bench improvement vs K2.6: +10.4% (vendor)
  • MCP Mark Verified improvement vs K2.6: +11.4% (vendor)
  • SWE Marathon improvement vs K2.6: +76.2% (vendor)
  • Reasoning-token use vs K2.6: 30% lower (vendor)
  • AIHackers repo eval: not verified (site-owned)
HighSpeed model ID exists at a higher token price; latency not independently measured here.Cheaper routine Kimi coding API lane when Kimi routing fits and 256K context is enough.Kimi K2.7 Code quickstart [archive], Kimi K2.7 Code pricing [archive], Kimi Code K2.7 release notes [archive], SWE-bench, Berkeley Function Calling Leaderboard2026-06-28
GLM-5.2Z.AIactive
Current Z.AI flagship coding model and supported-tool value lane.
1M$1.40 / 1M$4.40 / 1MZ.AI reports 62.1 on SWE-Bench Pro and 81.0 on Terminal-Bench 2.1.Supported-tool coding lane; BFCL score not imported.
  • Artificial Analysis Intelligence Index v4.1: 51 (independent)
  • SWE-Bench Pro: 62.1 (vendor)
  • Terminal-Bench 2.1: 81.0 (vendor)
Artificial Analysis flags higher output-token use; measure total cost per successful task.July value pick to test for supported coding-tool workflows; keep Opus/GPT for final arbitration until local evals pass.Z.AI GLM-5.2 overview [archive], Z.AI pricing [archive], Artificial Analysis: GLM-5.2 article [archive], Artificial Analysis Intelligence Index v4.1, SWE-bench, Berkeley Function Calling Leaderboard2026-06-28
Claude Sonnet 5Anthropicactive
Generally available across Claude plans, Claude Code, the Claude API, GitHub Copilot, and supported AWS paths.
1M$2.00 / 1M$10.00 / 1MAnthropic reports substantial coding and agentic gains over Sonnet 4.6; independent normalized results are pending.Available in Claude Code and the Claude API; adaptive thinking is on by default.
  • Cross-model benchmark evidence: vendor-reported; updated chart and system card preferred (vendor)
  • Artificial Analysis task cost: $1.53 per Intelligence Index task at max (independent)
  • AIHackers repo eval: not verified (site-owned)
No site-owned normalized latency result is verified.First Claude cost/performance test before Opus 5; escalate only when the premium pass changes the accepted result.Anthropic Claude Sonnet 5 launch [archive], Claude Sonnet 5 migration guide [archive], GitHub Copilot Claude Sonnet 5 launch [archive], Claude Sonnet 5 on AWS [archive], Artificial Analysis: Claude Opus 5 [archive]2026-07-25
Claude Opus 4.8Anthropichistorical
Still available, but superseded by Opus 5 for current premium comparisons.
1M$5.00 / 1M$25.00 / 1MHistorical premium Claude baseline; use Opus 5 for new task-level comparisons.Still available for pinned integrations; new Claude premium routing should test Opus 5.
  • Artificial Analysis Intelligence Index v4.1: 56 (independent)
  • Artificial Analysis output speed: 57.3 tokens/s (independent)
Artificial Analysis measured 57.3 output tokens/s; provider and workload latency vary.Historical premium baseline. Use Claude Opus 5 for current Claude premium routing.Claude models overview [archive], Claude API pricing [archive], Artificial Analysis: Claude Opus 4.8 [archive], Artificial Analysis Intelligence Index v4.1, LMArena leaderboard dataset, Berkeley Function Calling Leaderboard2026-07-25
GPT-5.5OpenAIactive
Generally available prior-generation OpenAI model retained for existing integrations and comparisons.
1.05M API; 400K Codex$5.00 / 1M$30.00 / 1Mnot verifiednot verifiednot verifiednot verifiedPrimary coding seat while ChatGPT/Codex limits fit the workload.OpenAI GPT-5.5 API model page, OpenAI GPT-5.5 ChatGPT limits, Artificial Analysis: GPT-5.5, LMArena leaderboard dataset2026-06-28

Kimi Code membership uses provider-managed tool IDs; the K3 and K2.7 rows describe explicit pay-as-you-go API models. Record the actual displayed backend model for tool evaluations.

Kimi K3 has early independent Artificial Analysis evidence and Moonshot launch-suite evidence, but no AIHackers-owned repository eval. Kimi K2.7 Code’s earlier release notes report improvements over K2.6 of 10.4% on Program-Bench, 11.4% on MCP Mark Verified, and 76.2% on SWE Marathon, plus 30% lower reasoning-token use. Keep both sets source-labeled.

Evaluation Checklist

  • Confirm the displayed backend model and client version.
  • Pin the repository commit and task.
  • Record weekly quota, five-hour window, and monthly credit state.
  • Run tests and review the final diff.
  • Record retries, latency, accepted completion, and human repair.
  • Re-run after a backend alias upgrade before comparing with older results.

Pricing and Checkout

Kimi’s live product page exposes plan purchase controls, but fetched public pages do not provide a stable, complete checkout table suitable for an evergreen price claim. Treat the signed-in checkout as the final source for plan price, renewal, promotion, tax, region, and cancellation terms.

Do not describe old haggling, first-month pricing, trial banners, or fixed benefit counts as current unless the live checkout and terms still show them.

Verdict

Choose Kimi Code when you want the official Kimi CLI or a membership-backed route into compatible coding agents. Choose the explicit K3 or K2.7 API when model identity, per-token accounting, multimodal API behavior, or reproducible evaluation matters more than membership convenience.

Sources


Last verified: July 18, 2026. Backend aliases, credits, rate windows, supported clients, checkout prices, model entitlements, and promotions can change independently.