LIVE
Model10/13Meta releases Muse Wonder AI model, claiming it can equal ChatGPT
More ›Submit for Promotion

Value Board — Cheapest Strong Models

Intelligence score vs. live price across 5 models: how much intelligence you get per dollar.

Updated 2026-10-11Score: Artificial Analysis · Price: OpenRouterBlended: input 3 : output 1
Min. intelligence
Sort by
#ModelIntelligenceBlended $Score / $Context
1GLM 5.3 Flashz-aiopen weights41.8$0.2370.150/0.500176.01.0M
2MiMo-V2.6-Proxiaomiopen weights46.3$0.5440.435/0.87085.21.1M
3GLM 5.3z-aiopen weights44.8$1.230.039/4.8036.41.0M
4Step 5 Previewstepfun43.7$1.431.00/2.7030.71.0M
5Muse Spark 1.3meta45.1$2.001.25/4.2522.61.0M

Free Models

Zero-priced on OpenRouter — free daily quota applies (20 req/min; 50–1000 req/day).

Ling 3.1 Flash
inclusionaictx 262Kscore 41.1
Inkling Small (free):free
thinkingmachinesctx 1.0M
Inkling (free):free
thinkingmachinesctx 1.0M
Lyria 3 Pro Preview
googlectx 1.0M
Lyria 3 Clip Preview
googlectx 1.0M
Nemotron 3.5 Lightning (free):free
nvidiactx 1.0M
Nemotron 3 Ultra (free):free
nvidiactx 1.0M
Dots3-Note Preview (free):free
dots-studioctx 512K
Apodex 1.1 Mini (free):free
apodexctx 262K
Laguna S 2.1 (free):free
poolsidectx 262K
Laguna XS 2.1 (free):free
poolsidectx 262K
Gemma 4 26B A4B (free):free
googlectx 262K

Free Agent / Coding Tool Quotas

Extracted from official pages by a deterministic script — no LLM involved. Changes are detected and flagged daily; entries marked “manual” are human-verified with a date.

Gemini CLIFree$0

60 requests/min and 1,000 requests/day with personal Google account

个人 Google 账号即可，含 Gemini 3 系列与 1M 上下文

Googleauto-checked 2026-10-11
GitHub CopilotFree$0

Free plans including Copilot Free and Copilot Student · Limited to 2000 completions per month on Copilot Free · Pro $10 USD per month · Pro+ $39 USD per month · Max $100 USD per month

免费档可直接用（含补全与聊天额度）；学生 / 教师 / 热门开源维护者免费享 Pro 功能

GitHubauto-checked 2026-10-11
CursorFree$0

Hobby For the tinkerer Free · No credit card required · Limited Agent requests · Access to Composer

Hobby 免费档无需信用卡；另有 Students 计划（cursor.com/students）

Anysphereauto-checked 2026-10-11
通义灵码Free$0

企业免费开通 个人免费使用

个人版免费使用、企业版免费开通（含编程智能体）

阿里云auto-checked 2026-10-11
WindsurfFree$0

Free $0：Light quota to code with agents、模型可选范围受限、无限 inline edits 与 Tab 补全

官方定价页为 JS 渲染，需无头浏览器核实（脚本只跟踪核实日期）

Cognitionmanual check 2026-10-11
Claude CodePaidPro/Max 订阅或 API

无免费档：Claude 聊天有 Free $0，但 Claude Code 需 Pro / Max / API

官方定价页只有聊天产品有 $0 免费档；曾出现「Claude Code 永久免费」的假条目

Anthropicmanual check 2026-10-11
Qwen CodeUnknown待确认

官方文档未列出固定免费额度（按量计费 / 需登录账号），待人工核实后补数字

避免像旧条目那样凭空写「完全免费」

阿里云manual check 2026-10-11
Recent changes
  • 2026-10-11newGemini CLI — 新增：60 requests/min and 1,000 requests/day with personal Google account
  • 2026-10-11newGitHub Copilot — 新增：Free plans including Copilot Free and Copilot Student · Limited to 2000 completi
  • 2026-10-11newCursor — 新增：Hobby For the tinkerer Free · No credit card required · Limited Agent requests ·
  • 2026-10-11new通义灵码 — 新增：企业免费开通 个人免费使用
  • 2026-10-11newWindsurf — 新增：Free $0：Light quota to code with agents、模型可选范围受限、无限 inline edits 与 Tab 补全
  • 2026-10-11newClaude Code — 新增：无免费档：Claude 聊天有 Free $0，但 Claude Code 需 Pro / Max / API
  • 2026-10-11newQwen Code — 新增：官方文档未列出固定免费额度（按量计费 / 需登录账号），待人工核实后补数字

Cheapest Provider per Model

Same model, different inference providers on OpenRouter. Cheaper endpoints are often quantized (fp8/fp4) — check the quant tag.

DeepSeek V4 Pro 0423score 3616 providerssave 94%
▼ Baidu $0.171/$0.341fp8
▲ Reka $1.05/$10.50
GPT-6 Astrascore 49.67 providerssave 92%
▼ OpenAI $5.00/$25.00unknown
▲ OpenAI $60.00/$300.00
gpt-oss-120bscore 11.624 providerssave 87%
▼ Venice $0.030/$0.150int4
▲ Cerebras $0.350/$0.750
GLM 5.3 Flashscore 41.833 providerssave 85%
▼ InferenceNet $0.070/$0.050fp4
▲ Reka $0.060/$1.60
DeepSeek V4 Flash 0423score 34.316 providerssave 84%
▼ StreamLake $0.083/$0.165fp8
▲ Cloudflare $0.440/$1.32
GPT-6 Lunascore 21.57 providerssave 75%
▼ OpenAI $0.050/$0.250unknown
▲ OpenAI $0.200/$1.00
gpt-oss-20bscore 911 providerssave 73%
▼ Darkbloom $0.018/$0.090fp8
▲ Groq $0.075/$0.300
Ling 3.0 Flash VLscore 24.62 providerssave 65%
▼ Novita $0.021/$0.062bf16
▲ DeepInfra $0.060/$0.180
Ling 3.0 Flashscore 20.12 providerssave 65%
▼ Novita $0.021/$0.063unknown
▲ DeepInfra $0.060/$0.180
GPT-5 Nanoscore 134 providerssave 55%
▼ OpenAI $0.025/$0.200unknown
▲ Azure $0.055/$0.440
Nemotron 3.5 Lightningscore 12.95 providerssave 52%
▼ Darkbloom $0.025/$0.120int4
▲ CoreWeave $0.070/$0.200
MiMo-V2.6-Flashscore 37.98 providerssave 51%
▼ Darkbloom $0.100/$0.280fp4
▲ Io Net $0.260/$0.410
Qwen3.5-9Bscore 11.26 providerssave 51%
▼ Darkbloom $0.080/$0.130fp4
▲ Together $0.170/$0.250
DeepSeek V4 Flash Vision Expscore 34.83 providerssave 51%
▼ DeepInfra $0.216/$0.647fp8
▲ SiliconFlow $0.440/$1.32
Claude Opus 5.5score 53.611 providerssave 50%
▼ Amazon Bedrock $4.00/$20.00unknown
▲ Anthropic $8.00/$40.00
Grok 4.7score 46.45 providerssave 50%
▼ xAI $2.00/$6.00unknown
▲ xAI $4.00/$12.00
Hy3score 25.35 providerssave 34%
▼ Tencent $0.132/$0.528fp8
▲ AtlasCloud $0.200/$0.800
Claude Haiku 5.5score 34.510 providerssave 9%
▼ Google $0.100/$0.500unknown
▲ Amazon Bedrock $0.110/$0.550
Claude Fable 5score 49.66 providerssave 9%
▼ Claude Platform on AWS $10.00/$50.00unknown
▲ Google $11.00/$55.00
GLM 4.7 Flashscore 14.93 providerssave 5%
▼ Venice $0.060/$0.400fp8
▲ Novita $0.070/$0.400

Method: intelligence score (Artificial Analysis Intelligence Index) divided by blended price (input 3 : output 1, USD per 1M tokens, OpenRouter list price). Data is collected deterministically every day; no model inference is involved.