ByteDance
Trae Pro
T
IDE assistant
Free AI IDE with paid Pro. Token-based since Feb 2026; routes to frontier models.
MODEL
Claude
PRICE
$10/mo
LIMIT
5000/1 month
CONTEXT
128K
LEADERBOARD
Bang-for-buck, ranked
19 AI subscriptions in the $20–25 band, scored on six factors and price-adjusted. Switch the view, re-weight by priority, change currency.
BALANCED WEIGHTS · INDEX = SCORE × ($20 / PRICE) · HIGHER IS BETTER →
PRESET
PRIORITY
What matters most to you?
PLANS · 19 COMPARED
What you're actually buying
Every subscription in the $20–25 band — chat assistants, IDE tools, coding subs and app builders. Prices are real (June 2026); switch currency in the toolbar.
GitHub
Copilot Pro
©
IDE assistant
Usage-based since Jun 2026 ($15 monthly AI Credits). Unlimited completions; lives in your IDE.
MODEL
Claude
PRICE
$10/mo
LIMIT
300/1 month
CONTEXT
200K
Tabnine
Tabnine Dev
T
IDE assistant
Privacy-focused completion + chat. Can run fully private / air-gapped models.
MODEL
Multiple + private
PRICE
$12/mo
LIMIT
1000/1 month
CONTEXT
64K
Mistral
Le Chat Pro
M
Chat assistant
EU frontier lab. Fast, cheap chat sub with web + code interpreter + Canvas.
MODEL
Mistral Large
PRICE
$15/mo
LIMIT
300/1 day
CONTEXT
128K
Z.ai
GLM Coding Plan
Z
Coding subscription
Cheapest Claude-Code-compatible plan. Lite ~80 prompts/5-hr; Pro/Max scale to ~1,600.
MODEL
GLM-4.7
PRICE
$18/mo
LIMIT
80/5 hrs
CONTEXT
200K
AWS
Amazon Q Dev Pro
q
IDE assistant
AWS-native dev assistant. Higher limits + IP indemnity; deep AWS integration.
MODEL
Amazon Q
PRICE
$19/mo
LIMIT
1000/1 month
CONTEXT
200K
Anthropic
Claude Pro
C
Chat assistant
Claude Code shares the Pro sub. Best-in-class agentic coding; 5-hr rolling + weekly cap. No image/video gen.
MODEL
Claude Opus 4
PRICE
$20/mo
LIMIT
45/5 hrs
CONTEXT
200K
OpenAI
ChatGPT Plus
O
Chat assistant
GPT-5.x + image/video gen. Generous caps on 5-hr windows; ~10 Deep Research/mo.
MODEL
GPT-5.x
PRICE
$20/mo
LIMIT
160/3 hrs
CONTEXT
256K
Google
Google AI Pro
G
Chat assistant
1M-token context + video gen + Deep Research + 2TB storage. Best if you live in Google Workspace.
MODEL
Gemini 3 Pro
PRICE
$20/mo
LIMIT
200/1 day
CONTEXT
1M
Perplexity
Perplexity Pro
P
AI search
Routes to frontier models. Unlimited Pro search + $5 API credit. No agent-connector support.
MODEL
Multi-model
PRICE
$20/mo
LIMIT
600/1 day
CONTEXT
128K
Anysphere
Cursor Pro
→
IDE assistant
AI-native IDE. $20 frontier-model usage; unlimited Tab; Auto mode effectively unmetered.
MODEL
Frontier + Auto
PRICE
$20/mo
LIMIT
500/1 month
CONTEXT
256K
Cognition
Devin / Windsurf Pro
D
IDE assistant
Windsurf is now Devin (Devin Desktop). Agentic IDE on SWE-1.6 + frontier models; quota-based.
MODEL
SWE-1.6 + frontier
PRICE
$20/mo
LIMIT
200/1 day
CONTEXT
200K
JetBrains
JetBrains AI Pro
J
IDE assistant
For JetBrains IDE users. 10 credits/mo ($1 each); deep IDE integration.
MODEL
Multiple
PRICE
$20/mo
LIMIT
300/1 month
CONTEXT
128K
Vercel
v0
v
App builder
React/Next UI generation. $20 monthly token credits; Figma import; v0 API.
MODEL
v0 models
PRICE
$20/mo
LIMIT
200/1 month
CONTEXT
128K
Lovable
Lovable
L
App builder
Prompt-to-app builder ("vibe coding"). 100 credits/mo (scales to 10k) + 5 daily; credits roll over.
MODEL
Hosted LLMs
PRICE
$25/mo
LIMIT
100/1 month
CONTEXT
128K
StackBlitz
bolt.new
B
App builder
Full-stack app builder in the browser. 10M tokens/mo, no daily cap, 1-month rollover.
MODEL
Hosted
PRICE
$25/mo
LIMIT
250/1 month
CONTEXT
128K
Replit
Replit Core
R
App builder
Cloud IDE + Agent. $25 monthly credits (no rollover) or $20 billed annually; up to 5 collaborators.
MODEL
Replit Agent
PRICE
$25/mo
LIMIT
200/1 month
CONTEXT
128K
xAI
SuperGrok
X
Chat assistant out of $20–25 band
No $20 tier — $30 is the entry. Live X data is its edge; trails on coding-agent benchmarks.
MODEL
Grok 4
PRICE
$30/mo
LIMIT
100/2 hrs
CONTEXT
256K
Alibaba
Qwen Coding Plan
Q
Coding subscription out of $20–25 band
Lite ($10) closed to new users Mar 2026 — only Pro ($50, 90k req/mo) is open. High-volume coding.
MODEL
qwen3-coder
PRICE
$50/mo
LIMIT
6000/5 hrs
CONTEXT
256K
COMPARE · ALL THE NUMBERS
Every plan, every metric
Click any header to sort. INDEX is price-adjusted (score × $20/price); SCORE is the raw weighted average. SCORE & INDEX follow the priority you pick above.
| PLAN | PRICE | CONTEXT | LIMIT | PERF | CODE | MULTI | TOOLS | CONN | SCORE | INDEX ▼ |
|---|---|---|---|---|---|---|---|---|---|---|
| Copilot Pro | $10 | 200K | 300 | 82 | 90 | 60 | 78 | 0 | 67.3 | 134.7 |
| Trae Pro | $10 | 128K | 5000 | 76 | 80 | 50 | 68 | 0 | 62.3 | 124.7 |
| Tabnine Dev | $12 | 64K | 1000 | 70 | 78 | 40 | 60 | 0 | 58.0 | 96.7 |
| Le Chat Pro | $15 | 128K | 300 | 78 | 72 | 70 | 76 | 28 | 70.0 | 93.3 |
| Google AI Pro | $20 | 1M | 200 | 90 | 85 | 95 | 86 | 56 | 84.7 | 84.7 |
| Claude Pro | $20 | 200K | 45 | 90 | 95 | 70 | 88 | 72 | 81.3 | 81.3 |
| ChatGPT Plus | $20 | 256K | 160 | 91 | 85 | 90 | 92 | 39 | 81.0 | 81.0 |
| GLM Coding Plan | $18 | 200K | 80 | 80 | 86 | 55 | 70 | 61 | 72.0 | 80.0 |
| Cursor Pro | $20 | 256K | 500 | 86 | 93 | 58 | 82 | 17 | 72.7 | 72.7 |
| Perplexity Pro | $20 | 128K | 600 | 84 | 70 | 78 | 94 | 0 | 71.0 | 71.0 |
| Amazon Q Dev Pro | $19 | 200K | 1000 | 78 | 82 | 45 | 80 | 0 | 64.2 | 67.5 |
| Devin / Windsurf Pro | $20 | 200K | 200 | 80 | 88 | 52 | 80 | 0 | 65.7 | 65.7 |
| v0 | $20 | 128K | 200 | 74 | 80 | 62 | 78 | 0 | 64.3 | 64.3 |
| JetBrains AI Pro | $20 | 128K | 300 | 76 | 80 | 45 | 70 | 0 | 60.8 | 60.8 |
| Replit Core | $25 | 128K | 200 | 74 | 84 | 50 | 82 | 0 | 63.7 | 50.9 |
| SuperGrok | $30 | 256K | 100 | 87 | 80 | 78 | 84 | 44 | 76.0 | 50.7 |
| bolt.new | $25 | 128K | 250 | 72 | 83 | 55 | 76 | 0 | 63.0 | 50.4 |
| Lovable | $25 | 128K | 100 | 72 | 82 | 60 | 75 | 0 | 61.8 | 49.5 |
| Qwen Coding Plan | $50 | 256K | 6000 | 82 | 88 | 60 | 72 | 56 | 76.3 | 30.5 |
CONNECTIVITY
Where each plan plugs in
Can your $20 sub actually be used inside these agents/IDEs — or do you need a separate API plan? App builders and IDE products are mostly standalone.
| PLAN | OpenCode | Claude Code | Hermes Agent | Cursor | Aider | Continue.dev | SCORE |
|---|---|---|---|---|---|---|---|
| Claude Pro | 72 | ||||||
| GLM Coding Plan | API | 61 | |||||
| Google AI Pro | 56 | ||||||
| Qwen Coding Plan | API | API | 56 | ||||
| SuperGrok | API | API | 44 | ||||
| ChatGPT Plus | API | API | API | 39 | |||
| Le Chat Pro | API | API | API | API | API | 28 | |
| Cursor Pro | 17 | ||||||
| Perplexity Pro | 0 | ||||||
| Trae Pro | 0 | ||||||
| Copilot Pro | 0 | ||||||
| Devin / Windsurf Pro | 0 | ||||||
| Tabnine Dev | 0 | ||||||
| Amazon Q Dev Pro | 0 | ||||||
| JetBrains AI Pro | 0 | ||||||
| Lovable | 0 | ||||||
| bolt.new | 0 | ||||||
| Replit Core | 0 | ||||||
| v0 | 0 |
NATIVE SUB WORKS API API KEY ONLY NOT SUPPORTED
METHODOLOGY
How the scores are built
Transparent by design. Prices and limits are sourced; factor scores are our calibrated estimates — submit a correction if something looks off.
Bang-for-buck index
Weighted average of the six factors below, multiplied by ($20 ÷ price). A $30 plan must therefore be 1.5× as strong on the factors to tie a $20 plan. Pick a priority preset to re-weight the factors.
Performance
Reasoning + general knowledge, calibrated against MMLU-Pro and reasoning-benchmark medians.
Usage limits
Log-normalized message / credit cap per window, plus a small bonus for heavier "deep work" runs (research, agent sessions, builds).
Connectivity
Sum across six connectors (OpenCode, Claude Code, Hermes, Cursor, Aider, Continue.dev): native = 3, sub works = 2, API-only = 1, none = 0.
Multimodal
Breadth across vision, voice, image generation and video generation.
Tool use
Built-in web search, code interpreter, file handling and canvas.
Coding
SWE-bench Verified / HumanEval composite for agentic and inline coding.
DATA TRUTH · JUNE 2026
Real / sourced: monthly price, usage limits, model names and connector support — verified against vendor pricing pages where possible. Calibrated estimates: the 0–100 factor scores (performance, usage limits, connectivity, multimodal, tool use, coding). These are our opinion, normalized against published benchmark medians — not vendor-reported numbers. Out-of-band plans (SuperGrok $30, Qwen Pro $50) are included for completeness and flagged. Not affiliated with any vendor.