subagent

Find the right AI model for any task.

curl -fsSL https://models.developersdigest.tech/install.sh | sh

paste this in terminal, or browse the leaderboard

Works with

OpenAI

Anthropic

Google

One command for any task

subagent classifies your task and picks the best model automatically.

Research

Deep web research, multi-source synthesis

subagent build "find 10 series A startups in philly"

Coding

Code generation, debugging, refactoring

subagent build "write a Redis caching layer in Go"

Writing

Blog posts, emails, marketing copy

subagent build "draft a launch announcement email"

Extraction

Parse documents, extract structured data

subagent build "extract contacts from this PDF"

Reasoning

Math, logic, complex multi-step problems

subagent build "prove this algorithm is O(n log n)"

Realtime

Chat, autocomplete, low-latency tasks

subagent build "autocomplete search suggestions"

Total Models

346

via OpenRouter

With Benchmarks

AA intelligence scores

Providers

29 free models

Avg Price

$5.10

per million tokens

Intelligence Leaderboard

Source: Artificial Analysis

#	Model	Provider	Intelligence	Price / 1M	Context	Modality
1	Google: Gemini 3.1 Pro Preview	google	57	$7.00	1.0M	audio, file, image, text, video
2	Anthropic: Claude Opus 4.6	anthropic	46	$15.00	1.0M	text, image
3	Anthropic: Claude Sonnet 4.6	anthropic	44	$9.00	1.0M	text, image
4	MiniMax: MiniMax M2.5	minimax	42	$0.75	197K	text
5	xAI: Grok 4	x-ai	42	$9.00	256K	image, text
6	Z.ai: GLM 5	z-ai	41	$1.68	203K	text
7	Qwen: Qwen3.5 397B A17B	qwen	40	$1.36	262K	text, image, video
8	Qwen: Qwen3 Max Thinking	qwen	40	$2.34	262K	text
9	StepFun: Step 3.5 Flash (free)	stepfun	38	Free	256K	text
10	StepFun: Step 3.5 Flash	stepfun	38	$0.20	256K	text
11	OpenAI: o3	openai	38	$5.00	200K	image, text, file
12	Qwen: Qwen3.5-27B	qwen	37	$0.88	262K	text, image, video
13	MoonshotAI: Kimi K2.5	moonshotai	37	$1.33	262K	text, image
14	Qwen: Qwen3.5-122B-A10B	qwen	36	$1.17	262K	text, image, video
15	Kwaipilot: KAT-Coder-Pro V1	kwaipilot	36	$0.52	256K	text
16	Google: Gemini 2.5 Pro	google	35	$5.63	1.0M	text, image, file, audio, video
17	Google: Gemini 3.1 Flash Lite Preview	google	34	$0.88	1.0M	text, image, video, file, audio
18	OpenAI: GPT-5.2	openai	34	$7.88	400K	file, image, text
19	DeepSeek: DeepSeek V3.2 Speciale	deepseek	34	$0.80	164K	text
20	Inception: Mercury 2	inception	33	$0.50	128K	text

State of the Art

Smartest

1Gemini 3.1 Pro Preview

57 pts

2GPT-5.4 (xhigh)

57 pts

3GPT-5.3 Codex (xhigh)

54 pts

Fastest

1Mercury 2

627 tok/s

2Granite 4.0 H Small

372 tok/s

3Gemini 2.5 Flash-Lite (Sep)

337 tok/s

Best Value

1Qwen3.5 9B

290.9 pts/$

2MiMo-V2-Flash (Feb 2026)

273.3 pts/$

3gpt-oss-20B (high)

266.7 pts/$

Cheapest

1Gemma 3n E4B

$0.03/M

2LFM2 24B A2B

$0.05/M

3Nova Micro

$0.06/M

Lowest Latency

1Ministral 3 3B

0.4s

2LFM2 24B A2B

0.5s

3Ministral 3 8B

0.5s

API Reference

All endpoints return JSON. No auth required.

All models with benchmarks, sorted by intelligenceGET

curl "https://models.developersdigest.tech/api/openrouter?benchmarks=true&sort=intelligence&limit=10"

# Response: { count, total, hash, models: [{ id, name, creator, intelligence, blendedPrice, contextLength, ... }] }

Filter by provider and priceGET

curl "https://models.developersdigest.tech/api/openrouter?creator=openai&max_price=5&sort=value"

# Params: creator, modality, max_price, min_context, sort (intelligence|price|context|newest|value), limit, benchmarks

Task-based recommendationsGET

curl "https://models.developersdigest.tech/api/recommend?task=find+10+series+A+startups+in+philadelphia"

# Response: { task, classification, weights, tools, recommendations: [{ name, composite, qualityScore, ... }] }

SOTA rankingsGET

curl "https://models.developersdigest.tech/api/sota"

# Response: { byIntelligence, bySpeed, byValue, cheapest, lowestLatency }

Model changelog (new/removed models)GET

curl "https://models.developersdigest.tech/api/changes"

# Response: { changes: [{ added: [...], removed: [...], timestamp }] }
# Synced daily via Vercel cron at 8am UTC

CLI

Install

curl -fsSL https://models.developersdigest.tech/install.sh | sh

subagent scrape              # fetch latest model data
subagent build "your task"   # get model recommendations
subagent harness             # detect available runtimes
subagent bench "your task"   # benchmark across runtimes