Practical AI

AI Radar: your monthly model and tool briefing

The best model changes every few months, and so do the apps around it. We track what changed, re-test your chosen tools on your tasks and tell you each month what to switch, keep or avoid.

Start your monthly briefing
Each month a new model version takes the top spot on the eagle’s board. The eagle re-tests it on the team’s tasks, moves the recommendation and files a monthly briefing.

The best model keeps changing.

Who held the top spot, month by month.

The #1 changed 15 times in 24 months

Today Sep 2026

Anthropic’s Claude Opus 5.5 is #1. A team choosing now would pick Claude.

Coding now: Claude Opus 5.5 in Claude Code (Anthropic), since 23 Sep.

GPT-6 Astra in Codex (OpenAI) led 5 to 23 Sep.

Also new in Sep 2026: Claude Fable 5.1, GPT-6 Sol, GPT-6 Luna, Claude Sonnet 5.5, GPT-6.1 Sol.

Best at coding:Not yet measuredAnthropicGoogleOpenAI

Short leads:Google Gemini 3 Pro, 16 to 26 Nov 2025OpenAI GPT-5.6 Sol in Codex, 12 to 16 Jul 2026Moonshot AI Kimi K3, 16 to 27 Jul 2026Alibaba Qwen3.8 Max, 1 to 2 Sep 2026OpenAI GPT-6 Astra, 5 to 23 Sep 2026

See the data
MonthIndependent tests: #1People’s votes: #1Best at coding
Oct 2024o1-preview OpenAIChatGPT-4o OpenAI, tiedNot yet measured
Nov 2024o1-preview OpenAIGemini Exp 1121 Google, tiedNot yet measured
Dec 2024o1-preview OpenAIGemini Exp 1206 Google, tiedClaude 3.5 Sonnet Anthropic
Jan 2025o1 OpenAIGemini 2.0 Flash Google, tiedClaude 3.5 Sonnet Anthropic
Feb 2025o3-mini OpenAIGrok 3 (early) xAIClaude 3.7 Sonnet Anthropic
Mar 2025o3-mini OpenAIGemini 2.5 Pro GoogleClaude 3.7 Sonnet Anthropic
Apr 2025o4-mini OpenAIGemini 2.5 Pro GoogleClaude 3.7 Sonnet Anthropic
May 2025o4-mini OpenAIGemini 2.5 Pro Google, tiedGemini 2.5 Pro Google
Jun 2025Gemini 2.5 Pro GoogleGemini 2.5 Pro GoogleGemini 2.5 Pro Google
Jul 2025Grok 4 xAIGemini 2.5 Pro Google, tiedGemini 2.5 Pro Google
Aug 2025GPT-5 OpenAIGemini 2.5 Pro Google, tiedGPT-5 OpenAI
Sep 2025GPT-5 OpenAIGemini 2.5 Pro Google, tiedGPT-5 OpenAI
Oct 2025GPT-5 Codex OpenAI, tiedGemini 2.5 Pro Google, tiedGPT-5 OpenAI
Nov 2025Gemini 3 Pro Preview GoogleGemini 3 Pro Google, tiedClaude Opus 4.5 AnthropicGemini 3 Pro (Google) 16 to 26 Nov
Dec 2025Gemini 3 Pro Preview GoogleGemini 3 Pro Google, tiedClaude Opus 4.5 Anthropic
Jan 2026GPT-5.2 OpenAIGemini 3 Pro GoogleClaude Opus 4.5 Anthropic
Feb 2026Gemini 3.1 Pro Preview GoogleClaude Opus 4.6 Anthropic, tiedClaude Opus 4.6 Anthropic
Mar 2026Gemini 3.1 Pro Preview Google, tiedClaude Opus 4.6 Anthropic, tiedClaude Opus 4.6 Anthropic
Apr 2026GPT-5.5 OpenAIClaude Opus 4.7 Anthropic, tiedClaude Opus 4.7 Anthropic
May 2026Claude Opus 4.8 AnthropicClaude Opus 4.6 Anthropic, tiedClaude Opus 4.7 Anthropic
Jun 2026Claude Fable 5 AnthropicClaude Fable 5 Anthropic, tiedClaude Fable 5 Anthropic
Jul 2026Claude Opus 5 AnthropicClaude Fable 5 Anthropic, tiedClaude Opus 5 AnthropicGPT-5.6 Sol in Codex (OpenAI) 12 to 16 JulKimi K3 (Moonshot AI) 16 to 27 Jul
Aug 2026Claude Opus 5 AnthropicClaude Fable 5 Anthropic, tiedClaude Opus 5 Anthropic
Sep 2026Claude Opus 5.5 AnthropicClaude Opus 5.5 Anthropic, tiedClaude Opus 5.5 AnthropicQwen3.8 Max (Alibaba) 1 to 2 SepGPT-6 Astra (OpenAI) 5 to 23 Sep
  • A company’s best-ranked model is not always its newest.
  • Independent tests: the Artificial Analysis Intelligence Index. Independent: Artificial Analysis runs every evaluation itself rather than relying on vendor-reported numbers, and publishes its methodology.
  • People’s votes: the LMArena / Arena text leaderboard (style-controlled default view). Measures preference in chat, not task success; top models are usually statistically tied.
  • Each line is a company’s best model at each month end, placed by its rank on that leaderboard. Ranks, not scores, because both leaderboards changed their scoring during these two years.
  • Best at coding: the #1 on LMArena’s WebDev Arena, then its Code Arena from November 2025.
  • Chosen for credibility and relevance, not for how often the top changes: the #1 model changed 15 times on Artificial Analysis (8 changes of company) and 11 times on LMArena (4).
Sources (59)

Leaderboards

Vendor sources

Computer use grew up fast.

One desktop task in seven in 2024. Ahead of people by December 2025.

72.6%

Simular’s Agent S3, using Claude Opus 4.5 and GPT-5 (best of 10 tries)

Dec 2025OSWorld Verified, up to 100 steps, several tries per taskSource: Simular's computer use agent outperforms humans

Not quite like for like
  • Until July 2025, makers reported their own OSWorld scores, with different screen inputs and step limits (15 to 200 steps).
  • OSWorld Verified, from 28 July 2025, fixed more than 300 tasks and has kept changing. A maker’s test set-up moves scores too: Anthropic re-scored Claude Opus 4.7 from 78.0% to 82.8%.
  • Agents built around a model, or that keep the best of several tries, are not like for like with one model’s single run.
  • OSWorld 2.0, from 26 June 2026, is a different and much harder test: 108 long tasks, up to 500 steps. It has its own line, which counts only finished tasks.
Sources (38)

Leaderboards

Vendor sources

Research

How AI connects keeps changing.

Seven ways in. Open one.

Model Context Protocol (MCP)

Good for
One standard for many systems.
Watch for
Poisoned tools, over-wide permissions.
Evidence

MCP lets one server work with many AI assistants. But many servers can flood the model with tool lists: Anthropic’s example fell from 150,000 to 2,000 tokens by calling tools from code. The standard’s guidance and researchers warn of hidden instructions in tool descriptions, servers tricked into misusing access, sign-in tokens passed on unchecked, untrusted local servers and over-wide permissions.

  1. Nov 2024Anthropic open-sources the Model Context Protocol
  2. Mar 2025OpenAI says it will support MCP, available that day in the Agents SDK
  3. Mar 2025MCP update adds standard sign-in, a new way to connect over the web, and labels for how each tool behaves
  4. Apr 2025Invariant Labs reveals ‘tool poisoning’ attacks: hidden instructions in MCP tool descriptions
  5. Apr 2025Official GitHub MCP Server in public preview
  6. Apr 2025Google says it will add MCP support to Gemini models and SDK

Ratings are our view; dates are sourced.

Sources (65)

Vendor sources

Research

Your monthly briefing.

One dated page.

October briefingSample, fictional firm
  • Switch
    Month-end commentary

    Your second assistant now handles the 12-tab workbook.

  • Keep
    Email on Microsoft 365

    Rivals improved, but not enough.

  • Watch
    A computer-use preview

    Fine for the supplier portal, not finance.

  • Avoid
    A free summariser extension

    It reads every site your staff visit.

Security note. One free tier now keeps chats longer. Tell staff.

How it runs.

  1. Track

    Models, prices and data terms.

  2. Re-test

    Your tasks, on what changed.

  3. Brief

    Switch, keep, watch or avoid.

  4. Talk

    Quarterly, or sooner if urgent.

    Back to track, every month

What you receive

  • Monthly briefing
  • Dated re-test log
  • Same-week security alerts
  • Annual toolset review

Why have a security team do this?

“Every new AI feature is another security question.”

Each month we check data handling, retention and permissions, not just capability.

How we do it

When a tool adds computer use or a new connector, you hear what it can reach first.

  • Data questions firstWhat could leak is settled before switch-on.
  • Access set up properlyWe configure and test it ourselves.
  • Verified with your peopleSessions in plain English, not slide decks.

Related Practical AI services

All Practical AI services
Which tool is best this month?

We keep watch and tell you when it matters.

Start your monthly briefing