---
title: "Reasoning effort — a pricing dimension, not a τ — Undominated"
canonical: https://undominated.ai/fr/effort/
description: "85 of 405 models cannot disable thinking (as of 2026-08-28). No reasoning-tax point estimate."
---

# Reasoning effort — a pricing dimension, not a τ — Undominated

> 85 of 405 models cannot disable thinking (as of 2026-08-28). No reasoning-tax point estimate.

# Reasoning effort

Max effort is a pricing dimension. We report what vendors publish. We do not ship a reasoning-tax τ point estimate — that needs first-party probes with median and IQR, n≥15.

As of 2026-08-28, **85** of 405 publishable models (21%) mark reasoning as mandatory — thinking cannot be turned off, so the billed completion rate is a cost floor, not an option. Condition: vendor reasoning.mandatory on the OpenRouter-derived catalogue; FakeProvider and Stealth excluded. No τ.

 85/405 cannot disable thinking
 274 reasoning supported
 189 supported and optional

Effort ladders below are vendor labels (low / high / xhigh / max). An Elo without its effort level is not a fact — see methodology.

| Model | Default | Efforts |
| --- | --- | --- |
| AionLabs: Aion-2.0 | — | — |
| AionLabs: Aion-3.0 | — | — |
| AionLabs: Aion-3.0-Mini | — | — |
| AllenAI: Olmo 3 32B Think | — | — |
| Anthropic: Claude Fable 5 | high | max, xhigh, high, medium, low |
| Anthropic: Claude Fable 5 (batch) | high | max, xhigh, high, medium, low |
| Arcee AI: Trinity Large Thinking | — | — |
| DeepSeek: R1 | — | — |
| DeepSeek: R1 0528 | — | — |
| Google: Gemini 2.5 Pro | — | — |
| Google: Gemini 2.5 Pro (batch) | — | — |
| Google: Gemini 2.5 Pro Preview 05-06 | — | — |
| Google: Gemini 2.5 Pro Preview 06-05 | — | — |
| Google: Gemini 3.1 Pro Preview | medium | high, medium, low |
| Google: Gemini 3.1 Pro Preview (batch) | medium | high, medium, low |
| Google: Gemini 3.1 Pro Preview Custom Tools | medium | high, medium, low |
| Google: Gemini 3.5 Flash | medium | high, medium, low, minimal |
| Google: Gemini 3.5 Flash (batch) | medium | high, medium, low, minimal |
| Google: Gemini 3.5 Flash Lite | minimal | high, medium, low, minimal |
| Google: Gemini 3.5 Flash Lite (batch) | minimal | high, medium, low, minimal |
| Google: Gemini 3.6 Flash | medium | high, medium, low, minimal |
| Google: Gemini 3.6 Flash (batch) | medium | high, medium, low, minimal |
| Google: Gemini 3.7 Flash | medium | high, medium, low |
| Google: Gemini 3.7 Flash (batch) | medium | high, medium, low |
| Google: Nano Banana Pro (Gemini 3 Pro Image Preview) | — | — |
| Google: Nano Banana Pro (Gemini 3 Pro Image) | — | — |
| LiquidAI: LFM2.5-2.6B (free) | — | — |
| Meta: Muse Glimmer 30B | medium | xhigh, high, medium, low |
| Meta: Muse Spark 1.1 | medium | xhigh, high, medium, low, minimal |
| Meta: Muse Spark 1.2 | medium | xhigh, high, medium, low, minimal |
| Meta: Muse Spark 1.2 Contributor | medium | xhigh, high, medium, low, minimal |
| MiniMax: MiniMax M2 | — | — |
| MiniMax: MiniMax M2.1 | — | — |
| MiniMax: MiniMax M2.5 | — | — |
| MiniMax: MiniMax M2.7 | — | — |
| MiniMax: MiniMax M2.7 (free) | — | — |
| MoonshotAI: Kimi K2 Thinking | — | — |
| MoonshotAI: Kimi K2.7 Code | — | — |
| MoonshotAI: Kimi K2.7 Code (batch) | — | — |
| OpenAI: GPT-5 | medium | high, medium, low, minimal |
| OpenAI: GPT-5 (batch) | medium | high, medium, low, minimal |
| OpenAI: GPT-5 Codex (batch) | — | — |
| OpenAI: GPT-5 Image | — | — |
| OpenAI: GPT-5 Image Mini | — | — |
| OpenAI: GPT-5 Mini | medium | high, medium, low, minimal |
| OpenAI: GPT-5 Mini (batch) | medium | high, medium, low, minimal |
| OpenAI: GPT-5 Nano | medium | high, medium, low, minimal |
| OpenAI: GPT-5 Nano (batch) | medium | high, medium, low, minimal |
| OpenAI: GPT-5 Pro | high | high |
| OpenAI: GPT-5 Pro (batch) | high | high |
| OpenAI: GPT-5.1-Codex | medium | high, medium, low |
| OpenAI: GPT-5.1-Codex-Max | medium | xhigh, high, medium, low |
| OpenAI: GPT-5.2 Pro | medium | xhigh, high, medium |
| OpenAI: GPT-5.2 Pro (batch) | medium | xhigh, high, medium |
| OpenAI: GPT-5.2-Codex | medium | xhigh, high, medium, low |
| OpenAI: GPT-5.4 Pro | medium | xhigh, high, medium |
| OpenAI: GPT-5.4 Pro (batch) | medium | xhigh, high, medium |
| OpenAI: GPT-5.5 Pro | medium | xhigh, high, medium |
| OpenAI: GPT-5.5 Pro (batch) | medium | xhigh, high, medium |
| OpenAI: gpt-oss-120b | medium | high, medium, low |
| OpenAI: gpt-oss-20b | medium | high, medium, low |
| OpenAI: gpt-oss-safeguard-20b | — | — |
| OpenAI: o3 Mini High | high | high |
| OpenAI: o3 Mini High (batch) | high | high |
| OpenAI: o4 Mini High | high | high |
| OpenAI: o4 Mini High (batch) | high | high |
| Perplexity: Sonar Pro Search | — | — |
| Qwen: Qwen3 235B A22B Thinking 2507 | — | — |
| Qwen: Qwen3 30B A3B Thinking 2507 | — | — |
| Qwen: Qwen3 Next 80B A3B Thinking | — | — |
| Qwen: Qwen3 VL 235B A22B Thinking | — | — |
| Qwen: Qwen3 VL 30B A3B Thinking | — | — |
| Qwen: Qwen3 VL 8B Thinking | — | — |
| Qwen: Qwen3.8 2.4T A95B | xhigh | xhigh, medium, low |
| Qwen: Qwen3.8 Max | xhigh | xhigh, high, medium, low, minimal |
| Reka Flash 3 | — | — |
| Sakana: Fugu Ultra | xhigh | max, xhigh, high |
| SpaceXAI: Grok 4.20 Multi-Agent | medium | xhigh, high, medium, low |
| SpaceXAI: Grok 4.5 | high | high, medium, low |
| SpaceXAI: Grok 4.6 | high | xhigh, high, medium, low |

Showing 80 of 85 mandatory models.