Reasoning AI models: they think before answering, and bill for it

Prices from official vendor price lists · October 2026
169 models that work through a problem step by step before answering. Output prices run from $0.059 to $180 per 1M tokens across the 145 models with a paid per-token price; 58 have a free tier and 76 ship open weights.
Reasoning
Services found: 169
#AI modelCapabilitiesAPI $/1M (in / out)Context windowPrice / moOutput $/1MOverviewVisit site
161

MiMo-V2.6-Pro-UltraSpeed

Xiaomi
Text & chatVisionAudio
$4.35 / $8.71M tokensAPI only8.7View
162

Hy4 preview

Tencent
Text & chatReasoningAdjustable thinking
$0.84 / $2.521M tokensAPI only2.52View
163

Fugu Max

Sakana
Text & chatReasoningAdjustable thinking
$2 / $6API only6View
164

Aion 3.5

Aion Labs
Text & chatReasoningPrompt caching
$3 / $6256K tokensAPI only6View
165

Aion 3.5 Mini

Aion Labs
Text & chatReasoningPrompt caching
$0.7 / $1.4256K tokensAPI only1.4View
166

Solar Mini 4

Upstage
Text & chatReasoningAdjustable thinking
$0.1 / $0.4512K tokensAPI only0.4View
167

Perceptron Mk1.5

Perceptron
Text & chatVisionAudio
$0.15 / $1.536.9K tokensAPI only1.5View
168

Ling 3.0 Flash VL

InclusionAI
Text & chatVisionVideo
$0.02 / $0.059256K tokensAPI only0.059View
169

Ternary Bonsai 2 27B

PrismML
Text & chatVisionReasoning
262.1K tokensAPI onlyView

Browse by need

Reasoning models169 modelsCoding assistants17 modelsVision and images156 modelsLong context 1M+82 modelsFree tier available80 modelsOpen weights88 models

A reasoning model works through a problem before it answers, generating hidden thinking tokens along the way. Those tokens bill as output, which is why reasoning costs more than plain generation and why the price spread here - $0.25 to $50 per million - matters more than it looks.

169 models expose reasoning, including every current flagship: Claude Fable 5.1, GPT-6 Astra, Claude Opus 5. Use them for maths, multi-step analysis and code review. For lighter tasks a reasoning model is an expensive way to be slow. Filter for Adjustable thinking as well - 72 models let you cap how long the model deliberates.

Thinking bills as outputThe hidden deliberation is charged at the output rate, which is why the same prompt costs more here.
Depth is adjustableMost current reasoning models let you cap how long they think, turning quality into a dial.
Not always betterOn extraction and classification a reasoning model is slower and dearer for identical results.

Models with this capability

The 169 models below all support this capability. Prices are input and output per 1M tokens from the vendor official price list; for dirhams multiply by 3.6725. Tap a card to open the model page.

Claude Fable 5.1
#1
Anthropic
in $10out $501M tokens
GPT-6 Astra
#2
OpenAI
in $10out $501.05M tokens
GPT-6.1 Sol
#3
OpenAI
in $2out $101.05M tokens
GPT-6 Sol
#4
OpenAI
in $2out $101.05M tokens
GPT-6 Luna
#5
OpenAI
in $0.1out $0.51.05M tokens
Claude Opus 5.5
#6
Anthropic
in $4out $201M tokens
Claude Sonnet 5.5
#7
Anthropic
in $2out $101M tokens
Grok 4.7
#8
xAI
in $2out $6500K tokens

How to choose

  • Use reasoning where the first plausible answer is often wrong: maths, analysis, code review, planning.
  • Prefer models with an adjustable depth control so you can tune cost per task.
  • Watch output limits — deliberation plus answer can hit the max output ceiling on long tasks.
  • Compare on your own hard cases, not on benchmark scores.

Common mistakes

Leaving thinking on maximumDefault depth is rarely optimal. Most tasks finish just as well at a lower setting.
Using reasoning for formattingReshaping text into JSON needs structured output, not deliberation.

Who this suits

AnalystsMulti-step financial and legal reasoning where a missed condition is expensive.
Engineering teamsCode review, migration planning and debugging across many files.
Research workSynthesis across conflicting sources, where the model must weigh rather than summarise.

Related filters

Adjustable thinking79 models with a reasoning depth dial
Frontier tier41 flagship models, the best each vendor sells
Code execution47 models that run code in a sandbox
Tool calling166 models that call your functions
All 239 modelsThe full rating with filters and prices

Frequently asked questions

Why do reasoning models cost more?

They generate hidden thinking tokens before the visible answer, and those tokens bill at the output rate. A hard question can produce several times more billed tokens than its answer suggests.

Can I limit how much a model thinks?

Yes on 72 of them. Claude, Grok, Qwen3 and DeepSeek V4 expose a depth control; DeepSeek switches thinking off entirely.

When is reasoning not worth it?

Classification, routing, extraction and summarising short text. A fast model does these as well for a tenth of the price.

Do reasoning tokens appear in the response?

No — you pay for them but do not see them. Some vendors expose a summary of the reasoning instead.

Is a reasoning model slower?

Noticeably. Expect seconds rather than milliseconds, which rules it out of autocomplete-style features.

Frequently searched

By capabilityReasoningCodingVisionLong context 1M+Free tierOpen weightsLow latencyRealtime voiceImage generation
By vendorAnthropic (13)OpenAI (30)Google DeepMind (17)xAI (8)Meta (6)DeepSeek (9)Mistral AI (11)Alibaba (Qwen) (34)Z.ai (12)Moonshot AI (5)
Popular modelsClaude Fable 5.1GPT-6 AstraClaude Opus 5Claude Fable 5GPT-5.6 SolGLM 5.3Grok 4.6Kimi K3GPT-5.6 Terra
By deliveryBatch APIPrompt cachingPriority tierFrontier tierProprietary APITool callingStructured JSONWeb searchMultilingual

Ratings 0

Avg. score 0

Rate this page

Help us improve