Agentic WeeklyGet The Weekly Verdict, freeSubscribe free
Menu

The Weekly Verdict · free email · every Monday

The week's AI releases, plus 5 ways to use them.

A free email every Monday: the week's few big AI releases, then 5 short use cases for the new models, each with a prompt you can paste.

Free. Double opt-in. One click to unsubscribe. We email a confirmation link first. Privacy

Newest model we track: Hy4 from Tencent, released Nov 4, $0.834 / $2.501 per 1M tokens in / out. Reported = read from official sources. Tested = we ran it. How we test

open weightsproprietarylicense not statedreleased in the last 7 days

Newest

The 6 most recently released models in our data, newest first. Bar = input price per 1M tokens, as stated by the vendor.

  1. Hy4NewTencent · license not stated$0.834in · $2.501 out
  2. Sarvam PrismNewSarvam AI · open weightsprice not published
  3. Hydrogen-1.0NewDeep Principle · open weightsprice not published
  4. LongCat-2.5NewLongCat (Meituan) · Promolicense not stated$0.30in · $1.20 outpromo price, list not stated
  5. Clover-35B-A3B-PreviewNewDiffuSpace · open weightsprice not published
  6. GPT-Rosalind-6NewOpenAI · proprietaryprice not published

New this weekAll 101 models

New this week

Models released in the last 7 days

  1. Hy4$0.834Tencent · · license not stated · input per 1M tokens
  2. Sarvam PrismPrice not publishedSarvam AI · · open weights
  3. Hydrogen-1.0Price not publishedDeep Principle · · open weights
  4. LongCat-2.5$0.30LongCat (Meituan) · · license not stated · input per 1M tokens
  5. Clover-35B-A3B-PreviewPrice not publishedDiffuSpace · · open weights

3 more. All models released this week

Cheapest

Lowest list (standing) input price per 1M tokens among the 50 models that state a USD price. Temporary launch or promo prices are not ranked; models running one carry a Promo tag. Not a quality ranking, and non-USD prices are left out.

  1. Claude Haiku 5.5Anthropic · proprietary$0.10in · $0.50 out
  2. GPT-6 LunaOpenAI · proprietary$0.10in · $0.50 out
  3. MiniMax M3.1 FlashMiniMax · open weights$0.15in · $0.60 out
  4. Mistral Small 4Mistral AI · open weights$0.15in · $0.60 out
  5. Ministral 3 14BMistral AI · open weights$0.20in · $0.20 out
  6. Mercury 3Inception · Promoproprietary$0.20in · $0.70 outpromo now $0.10 / $0.35

All 50 priced modelsBest LLM for coding

The wire

The 8 latest news briefs after the lead, each with a primary source and an evidence label.

  1. UNCUT (open source)Tool

    UNCUT 0.1 is a 7 MB headless coding-agent harness that never trims context or output

    Reported ·

  2. AnthropicTool

    Claude can now sign employees in to plugins itself: Claude-issued auth beta for Team and Enterprise

    Reported ·

  3. ggml-orgTool

    llama.cpp v0.9.0 adds FP8 E4M3 quantization, Mistral Large 4 and Chatterbox TTS

    Reported ·

  4. Sarvam AIModel

    Sarvam Prism: an open-weights decision model on Sarvam 105B that returns probabilities, not text

    Reported ·

  5. TencentModel

    Tencent releases Hy4 with vision and computer use, at $0.834 per million input tokens

    Reported ·

  6. LangGeniusTool

    Dify 1.19.0 adds nested workflow tracing, portable Workflow Tools and Skill commands in difyctl

    Reported ·

  7. Agent Autoloop (open source)Tool

    Agent Autoloop 1.0 runs Claude Code and Codex in the background, asking only when you must decide

    Reported ·

  8. Comfy OrgTool

    ComfyUI v0.44.0 adds MiniMax-H3 support, ProRes video saving, a VAE dtype node and new partner nodes

    Reported ·

All news briefsThis weekRSS

Test bench

Counts come straight from the public test logs. A smoke test is an install plus a short task, not a benchmark. How we test

  • 57public test logs, latest
  • 48passed
  • 6partly worked
  • 2failed
  • 1could not run
  • 43of 139 tools smoke-tested; the rest are Reported

Index

The rest of the site, in short lists. Each list links to the full page.

About Agentic Weekly

Agentic Weekly (theagenticweekly.com) is an independent weekly radar of new AI and agent tool releases, operated independently from Texas, US. It is a different site from theagenticleaderboard.com. About us · How we test · How we differ from other AI tool sites.