Skip to content
Projects

codertesla/ai-coding-deals has 141 GitHub stars and 6 forks. GitHot measured +1 stars in 24 hours and +3 in 7 days. Last checked 2026-11-18 03:36 UTC. Last push 2026-11-14.

How these counts are measured

Measured star growth

in 24h
+1
in 7 days
+3
in 15 days
Collecting history
in 1 month
Collecting history

Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.

About

Money-saving guide to AI coding tools: free tiers, discounts, referral credits & open-source alternatives. 省钱指南。

Developer Tools

README.md

Original

AI Coding Tools: Pricing & Free Tiers

Verified pricing, free quotas, referral credits, student perks, and zero-budget BYOK setups across 30+ AI coding IDEs, CLI agents, and LLM APIs.

English · 简体中文

GitHub stars Last commit Contributors Issues Tools tracked Last verified PRs welcome License

Featured · Zero-Cost · Agentic IDEs · Extensions · CLI Agents · Promos · Students · Matrix · Playbook · Sources

AI coding subscriptions, credit multipliers, and token quotas shift almost monthly. This reference tracks live subscription tiers, free-tier rate limits, referral credits, student packs, and low-cost Bring-Your-Own-Key (BYOK) routes across Cursor, Claude Code, GitHub Copilot, Cline, OpenCode, Devin, Google Antigravity, Kiro, DeepSeek V4.1 Flash, and Kimi K3—audited against official vendor documentation as of November 2026.

Featured AI Coding Deals & Credits

Ranked by cost-to-capability ratio. Expired promotions, paused signups, and enterprise-only tools are excluded from this shortlist and documented in their respective category tables below.

[!TIP] DeepSeek-V4.1-Flash (GA 2026-09-10; the snapshot now serving traffic is V4.1-Flash-1017): native multimodal. Two different model IDs—do not mix them up:

  • Official DeepSeek API: deepseek-flash (base https://api.deepseek.com). The retired aliases deepseek-v4-flash and deepseek-v4-flash-vision-exp are still accepted, but those models are retired: requests are served by V4.1-Flash-1017 and billed at Flash rates.
  • OpenCode Go: deepseek-v4.1-flash (base https://opencode.ai/inference/go/openai/v1/...). The official deepseek-flash name is not the model ID on this endpoint.

Official off-peak rates per 1M tokens: Flash $0.15 input / $0.60 output / $0.003 cache-hit; DeepSeek V4 Pro $0.66 / $1.98 / $0.022. Peak is 2× (Mon–Fri 01:00–04:00 & 06:00–10:00 UTC, excluding Chinese public holidays). Search and Fetch APIs bill $0.75 per 1,000 requests as of 2026-11-01 Beijing time (free before that). Lowest-cost access routes right now:

  1. Freebuff (free): ad-supported CLI, Web, and Desktop agent with zero API key setup.
  2. OpenCode Go ($10/mo): DeepSeek V4.1 Flash carries a standing $100/mo allowance on Go ($200 on Go Plus)—no code, no expiry date.
Tool Current Offer & Quota Mechanics Operational Verdict
OpenCode Go (Verified) Go: $10/mo flat—the $5 first-month promo ended 2026-08-24, so $10 is the only price now. Referral via this link credits $5 of Go usage to both parties. Model ID on Go is deepseek-v4.1-flash. DeepSeek V4.1 Flash carries a $100/mo allowance (5-hour window $20, weekly $50) at official peak/off-peak token rates; DeepSeek V4 Pro is capped at $15/mo. Tiers: Go Plus $40/mo (2× limits), Go Teams $20/seat/mo. Hitting a cap falls back to free models (Beam Free, Iceman Contributor Free) or, with Extra Usage enabled, to your pay-as-you-go balance. Also bundles Grok 4.7, Kimi K3 and K3.1 Pro, GLM-5.3 / 5.3-Flash, Qwen3.8 Max/Flash, MiniMax M3, GPT 6 Luna, Claude Haiku 5.5 and Gemini 4 Argon (~30 models). Highest-value paid route to DeepSeek V4.1 Flash: $100 of quota at DeepSeek's own per-token rates plus a portable API key.
Freebuff (Verified) Free, ad-supported coding agent across CLI, Web, Chat, Desktop and mobile (Android now, iOS soon) at freebuff.com (Web referral link). Priced at €0 and requires no API key or credit card. Routes DeepSeek Flash/Pro (legacy V4 IDs resolve to V4.1 Flash upstream), Kimi, and MiniMax. High-tier sessions enforce daily caps with automatic fallback to lighter models. Fastest zero-cost entry to DeepSeek V4.1 Flash; install via npm install -g freebuff and run immediately without managing keys.
CodeBuddy / WorkBuddy (Tencent CN, Verified) Register via the WorkBuddy invite link (WeChat login) for a 2,000-credit new-user gift (valid 1 month). Counted with in-app onboarding tasks, public materials cite a ceiling of 3,600 credits—roughly 90% of a Standard month's 4,000 credits (¥99/mo)—but the 1,600 beyond the invite gift is not itemized on the invite page, so read 3,600 as a cap rather than a guarantee. First paid upgrade within 30 days grants 500 bonus credits to both parties (valid 6 months). Official DeepSeek partner supporting V4.1 Flash and Kimi K3 (~1.62× credit multiplier). (HY3 free window ended 2026-07-22.) Largest regional signup credit grant; primary zero-cost testing ground for Kimi K3 and DeepSeek V4.1 Flash.
Tencent Cloud AI Model Square (Verified) Claim "New User Benefits Free Experience" in the top-right corner of Model Square: grants 1M free tokens per model across every listed language and multimodal model for 90 days (covering HY3, GLM, DeepSeek, Kimi Code, and more). Hunyuan HY3 also offers a Hy Token Plan starting at ¥28/mo for 35M tokens. Lowest-cost multi-model API backbone for BYOK clients; one endpoint feeds Aider, Cline, or Continue across multiple flagship models.
OpenAI Codex Bundled in ChatGPT Free across Web, CLI, and macOS/Windows desktop apps with no credit card required. Zero-friction access to OpenAI coding workflows for anyone holding an existing ChatGPT account.
Google Antigravity / Gemini High-ceiling free tier with free inline Tab completion; Google AI Pro unlocks higher 5-hour compute pools at $19.99/mo. Strongest free tier for full-stack web workflows, pairing parallel multi-agent execution with an integrated browser sandbox.
Grok / SuperGrok (xAI, Verified) 7-day SuperGrok free trial at grok.com/plans unlocks Grok 4.5+ and Grok Build (xAI's terminal coding agent) at full paid quotas. Requires a credit card; auto-renews at $30/mo unless canceled before day 7. Restricted to first-time subscribers. One-week full-quota benchmark window for Grok 4.5+ and Grok Build—cancel before day 7 to pay $0.
Cline (ClinePass) (Verified) ClinePass: $9.99/mo (cline.bot/cline-pass) bundles GLM 5.3 Flash / 5.3 / 5.2, Kimi K3, Kimi K2.7 Code / K2.6, DeepSeek V4 Pro / V4 Flash, MiniMax M3, MiMo V2.5, and Qwen3.7-Max / 3.7-Plus / 3.8-Max across Cline IDE, CLI, and Desktop. The plan runs inside Cline—there are no separate provider keys to export. (Earlier $1.99 launch rate has ended; check live checkout pricing.) Predictable $9.99/mo flat rate for open-weight coding models inside an autonomous VS Code and CLI agent.
Cursor Now sold as part of The X Subscription (one plan for Grok, Cursor, and X). Current tiers: Plus $8/mo · Premium $30/mo · Super $100/mo · Ultra $200/mo; teams $30 / $100 / $200 per seat; annual billing $80 / $300 / $1,000 / $2,000. Legacy Cursor plans (Pro $20, Pro+ $60, Teams Standard $40) are closed to new customers but keep working for existing subscribers. Plus is Grok-models-only; third-party models require Premium or above at API list price. Benchmark multi-model Agentic IDE on the new tier ladder; existing Pro subscribers should recalculate quota before switching.

Billing Rule of Thumb: Switching commercial AI coding subscriptions from monthly to annual billing cuts seat cost by 15%–20%. Spot a new promo or price shift? Submit a Pull Request.

Free & Zero-Budget AI Coding Stack ($0 / BYOK)

You can run a production-capable AI coding workflow on a $0 monthly software budget by combining open-source harnesses with free inference tiers:

  • Zero-Key Ad-Supported Agent: Freebuff provides a free CLI (npm install -g freebuff), Web, Desktop, and Chat harness routing to DeepSeek V4.1 Flash, Kimi, and MiniMax with no credit card or API key. (Operated by the Codebuff team as a standalone free product.)
  • Open-Source Clients + Local Models ($0): Pair Aider, Continue.dev, Cline, or OpenCode with local models via Ollama for zero token spend and complete local code privacy.
  • Included in Existing Free Accounts: OpenAI Codex (bundled with ChatGPT Free/Plus) and GitHub Copilot Free (2,000 completions/month).
  • Standalone IDE Free Tiers: Google Antigravity (free inline Tab completions + recurring 5-hour agent compute pools) and Cursor (Free tier, capped).
  • Free OpenAI-Compatible API Endpoint (NVIDIA NIM): NVIDIA NIM (base_url="https://integrate.api.nvidia.com/v1") issues a single API key covering a broad open-weight catalog at up to ~40 RPM (phone OTP required; rate-limited for prototyping, no SLA). Recommended baseline model: nvidia/nemotron-3-super-120b-a12b for fast, clean message.content responses.
  • 90-Day Multi-Model Trial Pool (Tencent Cloud): New users claiming the trial on Tencent Cloud AI Model Square receive 1M free tokens per model across all listed LLM and multimodal endpoints for 90 days, plus optional ¥28/mo for 35M tokens on Hunyuan HY3.

Agentic IDEs: Pricing & Quota Comparison

Standalone AI-native editors built for repository-wide context indexing, multi-file refactoring, and local or sandboxed command execution.

Tool Pricing (2026-11) Verified Discounts & Quota Rules Best Suited For Official Link
Cursor Free: capped
Plus: $8/mo
Premium: $30/mo
Super: $100/mo
Ultra: $200/mo
Teams: $30 / $100 / $200 per seat
Annual: $80 / $300 / $1,000 / $2,000
Now sold through The X Subscription; legacy Cursor plans are closed to new signups (Pro $20, Pro+ $60, Ultra $200, Teams Standard $40, Teams Premium $120 still run for existing subscribers). Paid tiers share one weekly usage pool across Cursor and Grok; Plus is limited to Grok 4.7 / 4.6 / 4.5 and Composer 2.5 at fixed settings, while third-party models start at Premium, billed at API list price from prepaid credit. The 50%-off referral program ended 2026-07-23. Weekly-quota multi-file editing, parallel background agents, and Grok 4.7 access inside one subscription shared with Grok and X. cursor.com
Devin Desktop (formerly Windsurf) Free (light quota, unlimited Tab)
Pro: $20/mo
Max: $200/mo
Teams: $80/mo base + $40/seat
$10 free AI credit upon connecting a Git repository during onboarding; Pro+ tiers include free SWE 1.6 access. (GLM-5.2 / Kimi K2.7 free promo ended 2026-07-05.) Offloading asynchronous, multi-step engineering tasks to Devin Cloud for automated PR generation. devin.ai
Google Antigravity Free tier included
AI Plus: $7.99/mo
AI Pro: $19.99/mo
AI Ultra: $99.99–$200/mo (cut from $250)
Free inline Tab completion across tiers; Ultra price cut from $249.99 to $200 in May 2026; agent compute pools replenish every 5 hours under a weekly ceiling. Full-stack web engineering with parallel multi-agent execution and live browser-in-the-loop debugging. antigravity.google
Kiro (AWS) Free: 50 credits
Pro: $20/mo (1,000 credits)
Pro+: $40/mo (2,000 credits)
Pro Max: $100/mo (5,000 credits)
Power: $200/mo (10,000 credits)
Annual billing discount available; overages billed at $0.04/credit. Credit multipliers: Auto 1.0×, Sonnet 1.3×, Opus 4.8 2.2×, Haiku 0.4×, GLM-5 0.5×, Qwen3 Coder Next 0.05×. Spec-driven architecture workflows and teams migrating from Amazon Q Developer IDE plugins. kiro.dev
Trae (ByteDance) Free (5K completions/mo, 2 concurrent tasks)
Pro: $20/mo ($20 usage)
Pro+: $60/mo ($60 usage, 3×)
Ultra: $200/mo ($200 usage, 10×)
Mainland China edition (trae.cn) is 100% free for individual developers with built-in Doubao and DeepSeek models. $20/mo Cursor-style IDE experience, or $0 domestic access via trae.cn. trae.ai
ZCode (Zhipu) New-user trial / GLM Coding Plan Multi-day new-user token trial available on signup. Coding Plan centers on GLM-5.3 and GLM-5.3-Flash (~3× effective quota vs. GLM-5.3). (1.5× multiplier promo ended 2026-06-30.) Dedicated desktop IDE tuned specifically for the Zhipu GLM-5.3 model family. zcode.z.ai
CodeBuddy (Tencent Intl) Free: 250 credits / 2 weeks
Pro: $9.95/mo (50% off $19.90 list) or $119.40/yr (1,000 credits/mo)
Team: $40/seat/mo
Recurring 250 free credits every 2 weeks; top-up packs from $9.95 per 1,000 credits. Official DeepSeek partner with V4.1 Flash support. International teams standardizing on Tencent Cloud toolchains at half the $20/mo IDE price point. codebuddy.ai
CodeBuddy / WorkBuddy (Tencent CN) Trial: Free (500 credits/mo)
Standard: ¥99/mo (2,000 + 2,000 bonus = 4,000 credits)
Advanced: ¥199/mo (4,000 + 5,000 = 9,000)
Flagship: ¥999/mo (20,000 + 30,000)
Buddy One VIP: ¥700/mo (list ¥1,399)
Recurring monthly: 30% off; annual: 44% off
WorkBuddy invite link grants a 2,000-credit new-user gift (1-month validity); with in-app onboarding tasks the cited ceiling is 3,600 total (~90% of one Standard month), though the extra 1,600 is not itemized publicly; 500 bonus credits each on first paid upgrade within 30 days. Supports Kimi K3 (~1.62×) and DeepSeek V4.1 Flash. Regional developers testing Kimi K3 or DeepSeek V4.1 Flash on signup credits before paying ¥99. workbuddy.cn / codebuddy.cn
Qoder CN (formerly Tongyi Lingma, Alibaba) Community: Free (capped, 2-week Pro trial + 300 credits)
Pro: ¥59/mo (2,000 credits)
Enterprise: ¥99–¥199/seat/mo
Rebranded and repriced on 2026-05-20, replacing its former unlimited free tier with strict completion and credit caps. Alibaba Cloud engineering teams switching between Qwen, GLM, and Kimi inside one IDE. cn.aliyun.com/product/lingma
Zed Core editor: Free
Zed Pro: Low-cost monthly tier
Open-source Rust editor; verified students receive 1 year of Zed Pro free including ~$10/mo in AI model credits. Sub-millisecond native editor responsiveness with local-first privacy and student credit grants. zed.dev

AI Coding Extensions for VS Code & JetBrains

Drop-in extensions that add inline autocomplete, codebase chat, and autonomous multi-file agents to your existing VS Code or JetBrains setup.

Tool Pricing (2026-11) Verified Discounts & Quota Rules Best Suited For Official Link
GitHub Copilot Free (2,000 completions/mo)
Pro: $10/mo (1,500 credits)
Pro+: $39/mo (7,000 credits)
Max: $100/mo (20,000 credits)
Business: $19/seat (1,900 credits)
Enterprise: $39/seat (3,900 credits)
Individual Pro/Pro+/Max signups reopened from 2026-06-17; Business/Enterprise card & PayPal signups reopened 2026-09-01. 1 credit = $0.01; inline completions and Next Edit remain unlimited on paid plans. Free Pro access for verified students. (Temporary 3,000/7,000 enterprise credit promo ended 2026-09-01.) Unlimited inline autocomplete at $10/mo and native GitHub PR/Actions integration. github.com/features/copilot
Continue.dev Core extension: $0 (Open-Source)
Team: ~$20/seat/mo
100% open-source; bring your own API key (BYOK) or connect local Ollama models at $0 software cost. Strict enterprise data governance, custom model routing, and air-gapped local inference. continue.dev
Cline Core extension: $0 (Open-Source BYOK)
ClinePass: $9.99/mo
BYOK and local models remain $0. ClinePass ($9.99/mo) bundles GLM 5.3 Flash/5.3/5.2, Kimi K3, DeepSeek V4 Pro / V4 Flash, MiniMax M3, MiMo, and Qwen3.7/3.8 across IDE, CLI, and Desktop (~2–5× standard API rate limits). Autonomous multi-step file editing and terminal execution inside VS Code with full BYOK control. cline.bot
Roo Code $0 (Open-Source BYOK) Zero software markup; pay only raw token rates on your connected LLM API keys. Custom role-based agent personas (Architect, Coder, Debugger) built on the Cline foundation. roocode.com
Augment Code Community: Free (capped)
Trial: 30K credits (card required)
Business: $100/mo flat (up to 50 seats, $100 usage included)
30K-credit evaluation trial; flat $100/mo Business base covers up to 50 seats with zero per-seat licensing fees. Small-to-midsize engineering teams navigating massive monorepos without paying per-seat taxes. augmentcode.com
Amazon Q Developer New IDE & paid signups blocked since 2026-05-15; End of Support on 2027-04-30 Existing subscribers retain access until 2027-04-30; AWS Console Q remains active. Do not start new IDE deployments here—migrate to Kiro. Legacy AWS IDE users planning a structured transition to kiro.dev. aws.amazon.com/q/developer
Supermaven Free tier (fast autocomplete)
Pro: $10/mo (1M token context, style adaptation, $5 chat credit)
Team: $10/seat/mo
Generous free autocomplete tier plus a 30-day Pro trial. Ultra-low-latency inline completions across 1M-token repository contexts. supermaven.com
Qodo (formerly CodiumAI) Developer: Free (30 org PR reviews/mo + 250 IDE/CLI credits)
Teams: $30/seat/mo (annual) or $38/mo
Free tier covers automated unit test generation and 30 monthly pull-request audits. Automated test-suite generation and pre-merge PR verification. qodo.ai

Legacy Free Tiers Discontinued (Enterprise-Only Transition)

These tools formerly offered popular individual free tiers before pivoting to enterprise-only licensing or removing bundled compute. Use the listed replacements instead of following outdated tutorials.

Tool Current Licensing Status Cutoff Timeline & Policy Change Recommended Individual Replacement
Sourcegraph Cody Enterprise-only (~$59/seat/mo, annual) Discontinued Free and Pro individual tiers in July 2025; directs individual developers to Amp. Amp (zero-markup pay-as-you-go API billing) or Continue.dev
Tabnine Enterprise-only ($39–$59/seat/mo, annual) Shut down individual free/paid plans and student access in April 2025. Cline, Continue.dev, or Aider
Warp Terminal free; bundled AI requires Build ($20/mo) Removed free bundled AI credits; free accounts are restricted to BYOK mode (OpenAI, Anthropic, Google APIs) with zero Warp-hosted compute. Cursor or open-source CLI agents (Aider, OpenCode) with your own API key

CLI & Terminal Coding Agents

Terminal-native agents that read your repository, edit files in place, run test suites, and commit directly to Git.

Tool Pricing (2026-11) Verified Discounts & Quota Rules Best Suited For Official Link
Freebuff Free (100% ad-supported, listed at €0)
Install: npm install -g freebuff
Free with no API key or card required; routes to DeepSeek V4.1 Flash (via Flash/Pro and legacy V4 IDs), Kimi, and MiniMax across CLI, Web, Chat, Desktop, and mobile (Android live, iOS pending) (Web referral link). Zero-budget terminal coding on DeepSeek V4.1 Flash with zero configuration. freebuff.com
Codebuff Subscriptions: $100/mo (1×), $200/mo (2.5×), $500/mo (7×)
Pay-as-you-go: $0.01/credit
500 free credits on signup (covers several hours of active repo work); cancel anytime. Commercial sibling to Freebuff. Deep repository indexing and high-autonomy terminal refactors without rate-limit throttling. codebuff.com
Claude Code Pro: $17/mo (annual) or $20/mo
Max: $100/mo (5×) or $200/mo (20×)
Team: $20–$100/seat/mo
~15% off on annual billing. Shares a rolling 5-hour usage window with Claude Web; Max tiers add dual weekly caps. Not included in Claude Free. Complex multi-file architecture refactors, deep debugging, and 1M-token reasoning tasks. claude.com/claude-code
Grok Build (xAI) Grok Free (chat only, rate-limited)
SuperGrok: $30/mo
7-day SuperGrok free trial at grok.com/plans unlocks Grok 4.5+ and Grok Build at paid quotas (card required; cancel before day 7 to avoid $30/mo charge). Benchmarking xAI frontier coding models in the terminal at $0 during the 7-day trial window. grok.com/plans
Kimi Code (Moonshot) Adagio: Free
Andante: ¥49/mo · Moderato: ¥99/mo ($19 intl)
Allegretto: ¥199/mo · Allegro: ¥699/mo
Phase-1 referral raffle has ended. Flagship Kimi K3 is live across Web, Code, and API (membership pricing). For a $0 Kimi K3 trial, use WorkBuddy signup credits first. Native Kimi K3 terminal agent workflows with shared Web/Code membership quotas. kimi.com/membership/pricing
OpenAI Codex Included in ChatGPT Free
Go: $8/mo · Plus: $20/mo
Pro: $100–$200/mo
$0 via ChatGPT Free (5-hour rolling window, no card required). Supports personal OpenAI API keys to bypass rolling windows (billed per token; excludes cloud sandbox). Developers in the OpenAI ecosystem using the macOS/Windows desktop app or cloud sandboxes. openai.com/codex
Gemini CLI / Antigravity CLI Free tier (high daily request ceiling) Generous complimentary daily quota with no upfront subscription required. Long-context repository queries and lightweight scripted CLI automation. antigravity.google
Ollama Cloud Free: $0 (1 concurrent cloud model)
Pro: $20/mo ($200/yr, 50× usage, 3 concurrent)
Max: $100/mo (10 concurrent)
Local inference is always unlimited and $0. Cloud GPU usage resets on 5-hour session and 7-day weekly windows. Unified local-first CLI workflow that bursts to large cloud models without juggling vendor API keys. ollama.com/pricing
Aider $0 (Open-Source BYOK / Local) 100% free software; pay only raw LLM API token rates or $0 with local Ollama models. Git-native terminal pair programming with automatic commit formatting across 75+ models. aider.chat
Amp (Sourcegraph) Pay-as-you-go (0% markup on raw API rates); $5 minimum deposit Ad-free; previously granted ~$10/day in hourly replenished free credits (paused or reduced for many accounts since May 2026). Unconstrained context windows and multi-model routing at raw upstream token cost. ampcode.com
Cline CLI Core CLI: $0 (Open-Source BYOK)
ClinePass: $9.99/mo (npm i -g cline)
BYOK remains $0; ClinePass ($9.99/mo) unlocks bundled GLM 5.3, Kimi K3, DeepSeek V4 Pro / V4 Flash, MiniMax, and Qwen3.8 quotas (details). Running Cline's autonomous agent loop headless in terminal pipelines on a fixed $9.99/mo open-weights plan. cline.bot
OpenCode Core CLI: $0 (Open-Source BYOK)
Go: $10/mo · Go Plus: $40/mo · Go Teams: $20/seat/mo
$5 Go referral credit for both parties via referral link. DeepSeek V4.1 Flash: $100/mo allowance (model ID deepseek-v4.1-flash; 5-hour window $20, weekly $50), DeepSeek V4 Pro $15/mo, plus Grok 4.7, Kimi K3 / K3.1 Pro, GLM-5.3, Qwen3.8, MiniMax M3, GPT 6 Luna, Claude Haiku 5.5 and ~30 models in total. Best $10/mo subscription for DeepSeek V4.1 Flash and open coding models inside a clean terminal UI. opencode.ai
Crush (Charm) $0 (Open-Source BYOK) Zero software cost; connect any custom API key for pay-as-you-go token billing. Crafted Terminal User Interface (TUI) ergonomics from Charm with flexible multi-model routing. github.com/charmbracelet/crush
Goose (Block) $0 (Open-Source BYOK / Local) Zero software cost; runs against local models or your own API keys. Extensible local automation built natively around Model Context Protocol (MCP) servers. block.github.io/goose
Qwen Code $0 (Open-Source) Zero licensing cost when paired with local or cloud Qwen coding models. High-efficiency terminal coding tailored to the Qwen3 Coder model family. github.com/QwenLM/qwen-code

Time-Limited AI Coding Promos

Active promotional windows with hard expiration dates or trial cutoffs. Automated weekly checks scan this table to flag expiring deals.

Tool Promotional Offer Expires Key Terms & Verification
Grok / SuperGrok (xAI) 7-day SuperGrok free trial — full paid quota on Grok 4.5+ and Grok Build via grok.com/plans TBD (no announced end date) Requires credit card; cancel before day 7 to avoid $30/mo auto-renewal. Limited to accounts with no prior subscription history.
CodeBuddy / WorkBuddy (Tencent CN) 2,000-credit new-user gift via the invite link; invitee's first paid upgrade within 30 days adds 500 credits to each side 2026-11-30 Gift requires registration plus one completed conversation/task before the cutoff; credits valid 1 month. Referrer earns 50 credits per activated invitee, plus 100 when that invitee is active on 3 days within their first week.

Archive of Recently Expired Promos (Removed from Active Table): Copilot Business/Enterprise 3,000/7,000 credit bump (2026-09-01); OpenCode Go 4× Flash allowance (2026-09-20, superseded by the standing $100/mo Go allowance); OpenCode Go $5 first month (2026-08-24); Cursor 50%-off referral program (2026-07-23); WorkBuddy HY3 free window (2026-07-22); Devin GLM-5.2 / Kimi K2.7 free window (2026-07-05); ZCode / GLM Coding Plan 1.5× multiplier (2026-06-30); Kimi Phase-1 referral raffle (ended 2026-09).

Free AI Coding Tools for Students (.edu)

Verified students and educators (.edu email or enrollment proof) can unlock commercial Pro tiers at $0. Stack all four verified offers below for a complete $0/year development environment.

Tool Verified Student Entitlement Verification Portal
Google Gemini & Antigravity 1 year of Google AI Pro ($0) — unlocks Gemini Pro and full Google Antigravity Pro compute pools gemini.google/students
Zed Pro 1 year of Zed Pro ($0) — includes ~$10/month in AI model credits Verify inside the Zed desktop client (zed.dev)
GitHub Copilot Pro Free Copilot Pro subscription bundled via the GitHub Student Developer Pack education.github.com/pack
JetBrains IDEs Free individual licenses across all JetBrains IDEs (including JetBrains AI Assistant access) while enrolled jetbrains.com/student
Cursor Regional student offers were tied to the legacy Cursor plans; after the move to The X Subscription, confirm current eligibility before applying cursor.com

AI Coding Tool Comparison & Decision Matrix

Match your workload and monthly budget directly to the highest-leverage stack:

Developer Profile & Workload Recommended Stack Why This Combination Wins
$0 Budget / Local-First Privacy Freebuff + Aider / Cline + Local Ollama Zero recurring subscription fees; Freebuff handles cloud V4.1 Flash tasks at $0 while Ollama keeps sensitive code local.
Heavy DeepSeek V4.1 Flash Usage Freebuff (free) → OpenCode Go ($10/mo) Start on Freebuff for zero-cost exploration; step up to OpenCode Go for a standing $100/mo Flash allowance (5-hour window $20, weekly $50) plus a portable API key. Go Plus at $40/mo doubles the limits.
Solo Developer / Best $10/mo Baseline GitHub Copilot Pro ($10/mo) Lowest fixed monthly cost for unlimited inline completions, Next Edit, and 1,500 monthly agent credits.
Solo Developer / Primary Agentic IDE Cursor Premium ($30/mo) or Kiro Pro ($20/mo) Cursor leads on interactive speed and multi-model breadth; Kiro is purpose-built for spec-driven architecture and AWS Q Developer migrations.
Large-Scale Refactoring & Architecture Cursor Premium ($30/mo) + Claude Code ($17–$20/mo) Use Cursor for sub-second inline edits and UI work; dispatch Claude Code in the terminal for deep multi-file refactors and 1M-token reasoning.
Full-Stack Web & Browser Debugging Google Antigravity (Free / $19.99 AI Pro) Runs parallel coding agents wired directly to an interactive browser sandbox for live DOM and network verification.
Existing ChatGPT Free / Plus User OpenAI Codex (macOS/Windows App & CLI) Zero incremental spend; activates cloud sandbox and local CLI agent capabilities already included in your ChatGPT plan.
Regional / Open-Weights Maximizer WorkBuddy (2,000 invite credits + onboarding tasks, up to a cited 3,600) + Tencent Model Square (1M tokens/model) + OpenCode Go / ClinePass ($9.99/mo) Reportedly covers roughly 90% of one Standard month of Kimi K3 and DeepSeek V4.1 Flash usage via WorkBuddy and Tencent Model Square before paying a cent.

7 Rules to Cut AI Coding & Token Spend

  1. Audit credit burn multipliers, not just sticker price. A $20/mo plan is rarely a flat rate: Kiro bills Opus 4.8 at 2.2× vs. Qwen3 Coder Next at 0.05× (a 44× spread), while Copilot, Antigravity, and Cursor deduct from credit or compute pools on frontier models.
  2. Route by task difficulty. Bind fast, low-cost models (DeepSeek V4.1 Flash, GLM-5.3 Flash, Haiku, Gemini Flash) to inline completions, unit tests, and boilerplate. Save Opus, Sonnet, or o-series reasoning models strictly for cross-module architecture and root-cause debugging.
  3. Exploit DeepSeek V4.1 Flash off-peak windows. Run daily coding through Freebuff (free) or OpenCode Go ($10/mo, $100 Flash allowance). When billing direct API tokens (deepseek-flash on the official API, deepseek-v4.1-flash on Go), batch heavy jobs outside Mon–Fri 01:00–04:00 and 06:00–10:00 UTC to cut token cost by 50%.
  4. Decouple the agent harness from the inference provider. Open-source clients (Aider, Cline, Continue.dev, OpenCode, Roo Code) let you swap underlying API providers in seconds—routing to Tencent Model Square (1M free tokens/model), NVIDIA NIM ($0), or local Ollama models with zero vendor lock-in.
  5. Chain new-user credit grants before subscribing. Claim WorkBuddy's 2,000 invite credits plus onboarding task credits (cited ceiling 3,600), Tencent Model Square's 90-day 1M-token pools, Codebuff's 500 signup credits, and SuperGrok's 7-day trial to benchmark real repository workloads before committing to a monthly seat.
  6. Switch proven tools to annual billing. Once an IDE or CLI agent survives a 30-day workflow test, switching from monthly to annual billing locks in a 15%–20% discount.
  7. Re-benchmark every quarter. Vendor quotas, free tiers, and model multipliers shift every 60–90 days. A quarterly 15-minute audit prevents paying legacy seat prices for commodities that are now free upstream.

Contributing

This guide stays useful because developers report price shifts, quota changes, and new promos as soon as they drop:

  • Spot an outdated price, expired promo, or missing tool? Open an issue or submit a Pull Request.
  • Required PR fields: Include the official vendor URL, the verification date, and any expiration date for time-limited promos.
  • Formatting & sync rules: See CONTRIBUTING.md for table schema and bilingual sync guidelines.

If this reference saved your team money on AI tooling or API spend, star the repository so more developers can find it.

Sources & Verification Log

All pricing tiers, token rates, and quota rules are cross-checked against official vendor documentation and changelogs (last full audit: November 2026), including DeepSeek API Docs, DeepSeek Pricing, DeepSeek V4.1 Flash Release Note, OpenCode Go Docs, ClinePass, GitHub Copilot Changelog, Cursor Referral Program Notice, and Amazon Q Developer End-of-Support Announcement.

Always confirm live terms on the linked vendor pricing page before checkout. This repository is an independent community project unaffiliated with any vendor; clearly marked referral links help fund ongoing verification at zero extra cost to you.