Fable class
Fable-class (frontier) equivalents
The frontier models, one step above Opus: GPT-6.2 Astra in Codex, Daybreak Blue in Codex, and GPT-5.6 Sol Max in Codex.
For the hardest problems, at the highest price.
AI model equivalents
You know the Claude lineup by heart, then you open Codex, Gemini or Grok and none of the names tell you which model plays the role of Opus. This page places every model of every coding agent CLI AgentsRoom runs on one scale, so each one reads in terms you already know.
Read live from the AgentsRoom model catalogue. Popularity as of Nov 20, 2026, 7:48 PM.
CLIs compared
19
Models placed
231
Launches counted, last 30 days
215,051
Claude is the yardstick because it is the lineup most developers know. Each line names the models of other vendors that play the same role today.
Fable class
The frontier models, one step above Opus: GPT-6.2 Astra in Codex, Daybreak Blue in Codex, and GPT-5.6 Sol Max in Codex.
For the hardest problems, at the highest price.
Opus class
The models that play the role of Claude Opus: GPT-6.2 Sol in Codex, Gemini 3.1 Pro (High) in Antigravity, Grok 4.7 in Grok, DeepSeek V4 Pro (latest) in Command Code, Kimi K3 in Command Code, GLM-5.3 in Command Code, Qwen 3.8 Max Prime in Command Code, and Gemini 3.8 Flash in Command Code.
The flagship most serious coding work runs on.
Sonnet class
The models that play the role of Claude Sonnet: GPT-6.1 Terra in Codex, GPT-6 Luna in Codex, Gemini 3.8 Flash (High) in Antigravity, GLM 5.3 in OpenCode, Qwen3.8 Max in OpenCode, Kimi K3.1 Pro in Kimi, DeepSeek V4 Pro in Pi, and Medium 3.5 in Mistral Vibe.
Balanced: strong enough for most features, cheaper than the flagship.
Haiku class
The models that play the role of Claude Haiku: Codex Spark in Codex, DeepSeek V4 Flash in oh-my-pi, Qwen3 Coder in Ollama, GLM 4.6 in Ollama, and Gemini 3.5 Flash Lite in Command Code.
Light and fast: renames, small fixes, summaries.
One row per CLI AgentsRoom runs, one column per tier. Current models come first; the older versions a list still offers are counted under them.
| CLI | FrontierFable class | Opus class | Sonnet class | Haiku class |
|---|---|---|---|---|
FableFable 5.2 | OpusOpus 5.5 +2 older versionsOpus 5, Opus 4.8 | SonnetSonnet 5.5 +1 older versionSonnet 5 | Haiku | |
GPT-6.2 AstraDaybreak BlueDaybreak RedGPT-5.6 Sol Max +1 older versionGPT-6 Astra | GPT-6.2 Sol +4 older versionsGPT-6.1 Sol, GPT-6 Sol, GPT-5.6 Sol, GPT-5.5 | GPT-6.1 TerraGPT-6 Luna +4 older versionsGPT-5.6 Terra, GPT-5.6 Luna, GPT-5.4, GPT-5.2 | Codex Spark +1 older versionGPT-5.4 Mini | |
| None | Gemini 4 Argon (Low)Gemini 4 Argon (Medium)Gemini 4 Argon (High)Gemini 4 Carbon (Low)Gemini 4 Carbon (Medium)Gemini 4 Carbon (High)Gemini 4 Hydrogen (Low)Gemini 4 Hydrogen (Medium)Gemini 4 Hydrogen (High)Gemini 4 Hydrogen (XHigh)Gemini 4 Neon (Low)Gemini 4 Neon (Medium)Gemini 4 Neon (High)Gemini 4 Neon (XHigh)Gemini 3.1 Pro (High)Claude Opus 5.5 (High)Claude Opus 5.5 (Medium) | Gemini 3.8 Flash (High)Gemini 3.8 Flash (Medium)Claude Sonnet 5.5 (High)Claude Sonnet 5.5 (Medium) +5 older versionsGemini 3.7 Flash (High), Gemini 3.7 Flash (Medium), Gemini 3.6 Flash (High), Gemini 3.6 Flash (Medium), Gemini 3.5 Flash (Medium) | None | |
| None | GPT-5.5Claude Opus | Claude SonnetKimi K3GLM 5.3Qwen3.8 Max | None | |
| None | Grok 4.7 | None | None | |
GPT-6 Astra | GPT-6.1 SolClaude Opus 5.5GPT-5.5Gemini 4 Argon +4 older versionsGPT-6 Sol, Claude Opus 5, GPT-5.6 Sol, Claude Opus 4.8 | Claude Sonnet 5.5GPT-6 LunaGPT-5.6 TerraGemini 3.7 FlashKimi K3 +3 older versionsClaude Sonnet 5, GPT-5.6 Luna, Claude Sonnet 4.6 | Claude Haiku 5.5MAI-Code-1.1-Flash +1 older versionClaude Haiku 4.5 | |
| None | Claude Opus 5.5GPT-6 Sol | Claude Sonnet 5Gemini FlashGLM 5.3 +1 older versionGLM 5.2 | DeepSeek V4 Flash | |
| None | None | Qwen3 Coder 480B | Gemma 3Qwen3 CoderGLM 4.6 | |
| None | None | Kimi K3.1 ProKimi K3.1 Pro HighSpeedKimi K3Kimi K3-256KKimi K2.8 PreviewK2.7 Code HighSpeed | None | |
| None | Claude Opus 5.5GPT-6 Sol | Claude Sonnet 5Gemini FlashGLM 5.3DeepSeek V4 Pro +1 older versionGLM 5.2 | None | |
| None | Claude OpusGPTCodexGemini | SWEClaude SonnetGLMKimiDeepSeek | Claude Haiku | |
| None | None | Medium 3.5 | None | |
Claude Fable 5.1Claude Fable 5GPT-6 AstraGemini 4 Argon | DeepSeek V4 Pro (latest)Kimi K3GLM-5.3MiMo V2.6 ProMiMo V2.6 Pro UltraSpeedQwen 3.8 Max PrimeQwen 3.8 Max 0902Qwen 3.8 MaxStep 5 PreviewTencent Hy4 PreviewLing 3.1 FlashClaude Opus 5.5Claude Opus 5GPT-6.1 SolGPT-5.5Gemini 3.8 FlashFugu UltraMuse Spark 1.3Muse Spark 1.3 ContributorMuse WonderGrok 4.7 +5 older versionsClaude Opus 4.8, Claude Opus 4.7, GPT-6 Sol, GPT-5.6 Sol, Grok 4.6 | DeepSeek V4 Flash Vision (exp)DeepSeek V4.1 FlashDeepSeek V4.1 Flash FastKimi K3.1 ProKimi K2.7 CodeKimi K2.7 Code HighSpeedGLM-5.3 FlashGLM-5.3 FlashXGLM-5.2 FastMiniMax M3.1 FlashMiniMax M3.1 Flash FreeMiniMax M3MiMo V2.6 FlashQwen 3.8 27BQwen 3.8 FlashQwen 3.6 Max PreviewTencent Hy3Nemotron 3.5 Super VLInklingInkling SmallMistral Large 4Claude Sonnet 5.5Claude Sonnet 5GPT-6 LunaGPT-5.6 TerraGPT-5.3 CodexMuse Spark Open +21 older versionsDeepSeek V4 Flash (latest), DeepSeek V4 Flash Fast, Kimi K2.6, Kimi K2.5, GLM-5.2, GLM-5.1, GLM-5, MiniMax M2.7, MiniMax M2.5, MiMo V2.5 Pro, Qwen 3.7 Max, Claude Sonnet 4.6, GPT-5.6 Luna, GPT-5.4, Gemini 3.7 Flash, Gemini 3.6 Flash, Gemini 3.5 Flash, Muse Spark 1.1, Muse Spark 1.2, Muse Spark 1.2 Contributor, Grok 4.5 | MiMo V2.5Qwen 3.8 Omni FlashQwen 3.7 PlusLongCat 2.0Step 3.7 FlashNemotron 3 UltraClaude Haiku 5.5GPT-5.4 MiniGemini 3.5 Flash Lite +4 older versionsQwen 3.7 Flash, Qwen 3.6 Plus, Claude Haiku 4.5, Gemini 3.1 Flash Lite | |
| None | Muse Spark 1.3Muse WonderMuse Wonder Full | Muse Ember +1 older versionMuse Spark 1.2 | None | |
| None | Claude Opus | Claude SonnetDeepSeek | GPT-4o | |
| None | MiniMax M3 | MiniMax M3.1 Flash | None | |
| None | GLM-5.3GLM-5.3GLM-5.3 | GLM-5.3 FlashGLM-5.3 FlashGLM-5.3 Flash | None | |
| None | None | Qwen3.8 Max +1 older versionQwen3.7 Max | Qwen3.7 PlusQwen3 Coder PlusQwen3 Coder Next | |
| None | Claude Opus 4.8 | Claude Sonnet 4.6 | None | |
| Picks the model for you on every request: no fixed tier. | ||||
| Picks the model for you on every request: no fixed tier. | ||||
| Picks the model for you on every request: no fixed tier. | ||||
| Picks the model for you on every request: no fixed tier. | ||||
| Picks the model for you on every request: no fixed tier. | ||||
| Picks the model for you on every request: no fixed tier. | ||||
A model family is not tied to its vendor's tool. Here is every family that more than one CLI ships today, with the name each CLI gives it.
10 CLIs
10 CLIs
7 CLIs
5 CLIs
5 CLIs
5 CLIs
5 CLIs
4 CLIs
4 CLIs
3 CLIs
3 CLIs
3 CLIs
2 CLIs
2 CLIs
2 CLIs
2 CLIs
The most launched models of the last 30 days among AgentsRoom users, read in tier terms. Equivalents tell you which models play the same role; launches tell you which ones people pick.
OpusOpus class
27.6% of launches
SonnetSonnet class
13% of launches
Opus 5.5Opus class
10.4% of launches
HaikuHaiku class
10.2% of launches
GPT-6.1 SolOpus class
10% of launches
The four approximate tiers are anchored on Claude: Fable, Opus, Sonnet and Haiku. Reviewed models use an explicit catalogue tier; other entries retain the existing family-based fallback. These tiers are guides, not guarantees of equal results.
Command Code and Muse Code use published coding benchmarks for reviewed models. The word Spark alone does not mean a lightweight model: Muse Spark 1.3 is in Opus class, while Codex Spark remains in Haiku class.
Open weights do not impose a tier ceiling. A reviewed model can reach Opus class based on its coding results. Older catalogue entries may keep a different fallback until they are reviewed; the model name or its price alone does not establish an equivalence.
Routers get no tier, because they pick a model on each request: GitHub Copilot (Auto), Freebuff (Freebuff default), Cursor (Default), Devin (Adaptive), Amp (Amp default), Hermes Agent (Default), Goose (Default), Qwen Code (Default), Kiro (Default), and Rovo Dev (Default). Neither does a family the table does not know yet: no equivalence is better than a wrong one.
A tier is not a benchmark score. Two models of the same tier still differ in price, speed, context window, usage limits and the tasks they do best. Use the table to know where to look, then try both on your own task.
Data: the model catalogue AgentsRoom serves to its desktop and mobile apps, read at every visit, and the tier rules the app applies under every model list. Popularity: real agent launches from AgentsRoom desktops that have not declined usage sharing, over the last 30 days, refreshed every ten minutes.
In AgentsRoom
AgentsRoom is a desktop command center for AI coding agents. Claude, Codex, GitHub Copilot CLI, Cursor and 20 other agent CLIs run side by side, each agent on the model you choose. The equivalences of this page are the ones the app shows you while you choose, and a few smart features help you size the model to the task.
Every model list (new agent, quick launch, backlog ticket, mobile app) tags each model with its counterpart in the CLI you know best, and sorts by power, price or popularity.
Model equivalents in the appAI Suggestion turns a one-line description into a complexity level (Quick, Standard, Complex, Extreme) and opens the agent on the cheapest model that fits. Adaptive Mode checks your draft and suggests a lighter model when the flagship is overkill.
AI Suggestion and Adaptive ModeWhen an account nears its limit, a quota rule can start new agents on another CLI model by model: Fable, Opus and Sonnet each go to their counterpart of the same tier, pre-filled from these tiers.
Quota rules and usage alertsPin a provider, a model and a reasoning effort on a backlog ticket, so the agent that runs it uses exactly that, and keep the flagship for the tickets that need it.
Backlog task boardA developer agent on a flagship model hands the test run to a QA agent on a lighter one over MCP, so the expensive model stays on the code.
Agent delegationIn an agent team, every step has its own role, CLI and model: the architect on a frontier model, the implementer on the Opus class, the reviewer wherever it pays off.
Agent teamsAn honest comparison of Claude Code and Codex CLI: pricing, speed, MCP, control. And why, after a few weeks, I stopped choosing between them.
Ollama is now a provider in AgentsRoom. Run local open-source models like Llama, Qwen, Gemma and DeepSeek next to cloud agents, and set a local-or-cloud dial on each agent, switchable mid-conversation.
205,366 agent launches over 28 days: Opus first, GPT-6.1 Sol second, and half of accounts on one model we suggest by default. What the numbers prove, and what they can't.