Skip to content
Giallo
IT

AI model comparisons

Claude, GPT, Gemini and the other models compared on prices, official benchmarks and real tasks, to choose the right one for each job.

46 guides in this category

Claude Opus 5.5: Official benchmarks and prices compared to Fable 5.1 and GPT-6 AstraClaude Opus 5.5: official benchmarks, prices and comparison with Fable 5.1 and GPT-6 AstraClaude Opus 5.5, released 22 September 2026: $4 input and $20 output per million tokens, like Fable 5.1. Official benchmarks versus GPT-6 Astra.
GPT-6 Sol vs Claude Opus 5.5: comparing official OpenAI and Anthropic benchmarksGPT-6 Sol vs Claude Opus 5.5: the official numbers and when to use whichGPT-6 Sol vs Claude Opus 5.5, released on the same day: OpenAI and Anthropic benchmarks, why they cannot be compared 1:1, pricing and use cases.
GPT-6 Sol vs Luna: the two OpenAI models with the same context and a different price twenty timesGPT-6 Sol vs Luna: which one to choose and for which jobsGPT-6 Sol vs Luna: the same 1.05 million token context window with a 20-fold price difference. Official benchmarks, cost examples and a decision table.
How the latency of AI models influences choice in the companyAI model latency comparedHow response speed affects AI model selection, when it is critical and how to measure it on your own use cases.
How to Choose the AI Model to Extract Data from Business DocumentsBest AI model for extracting data from documentsChoose an AI model for extracting invoice and document data: the required model tier and a test to identify usable options.
How to choose an AI model for high volume machiningAI models for high volume workHow to choose and design an AI system that processes hundreds of thousands of cases per month without blowing up costs.
How Price per Million Tokens Work in AI ModelsPrice per million tokens explainedWhat pricing per million tokens means, what a token represents, why output costs more and how to estimate actual spending.
Pattern of a two-model AI strategy: light and powerfulMixed model strategy: light and powerfulCombine a lightweight and a powerful AI model: handoff criteria, checks and how much you actually save.
How to read AI model benchmarks and why vendor numbers are not comparableHow to read AI benchmarksWhat AI model benchmarks really say, why vendor numbers are not comparable and how to choose without getting fooled.
API prices of Claude Haiku 5.5, Sonnet 5.5, Opus 5.5 and Fable 5.1, with 100K threshold of Haiku promptClaude model pricing: API rates and the cost of getting work doneUpdated Claude model prices: Haiku, Sonnet and Opus 5.5, Fable 5.1. API price list, 100K threshold and a transparent calculation to estimate the budget.
Choice of Claude Haiku 5.5, Sonnet 5.5, Opus 5.5 and Fable 5.1 models according to task and testsClaude models compared: a guide to choosingCompare Haiku, Sonnet and Opus 5.5 with Fable 5.1: prices, context, reasoning and tests. Choose Claude for the task and plan a verifiable migration.
How to choose the best AI model to programBest AI Model for ProgrammingChoose an AI model for software development: which benchmarks to examine, what they miss and how to test it on your own code.
How to Choose the Right AI Model to Analyze Long DocumentsBest AI Model for Long DocumentsChoose an AI model for contracts, specifications and long documents: context windows, costs and the pitfall of sending everything.
All official GPT-6 Astra benchmarks comparedGPT-6 Astra: all official benchmarksGPT-6 Astra benchmarks from OpenAI's official tables: reasoning, coding, science, cybersecurity and the methodology notes to read.
GPT-6 Astra: what is the new OpenAI model and what changesGPT-6 Astra: what it is, how to use it and its limitationsWhat GPT-6 Astra is: specifications, ChatGPT and API access, benchmarks it does not win and limitations stated by OpenAI.
GPT-6 Astra for programming: benchmark coding by comparisonGPT-6 Astra for developers: code, API and computer useGPT-6 Astra for developers: coding scores, API endpoints and tools, reasoning levels, long context and computer use.
Official GPT-6 Astra prices per million tokensGPT-6 Astra: pricing, real costs and when it is worthwhileGPT-6 Astra's official pricing, costs the price list leaves out and when the model is justified for a business.
What is ARC-AGI testing and what it measures in AI modelsARC-AGI: what it is and why it is talked aboutWhat ARC-AGI is, the abstract reasoning test designed to resist memorization, and how much it matters to a company.
What AI benchmarks to look for based on business use caseWhich AI benchmarks matter?What benchmarks to look at based on your use case: coding, agents, computer use, reasoning. And which to ignore.
Why benchmarks published by AI vendors are not comparable to each otherWhy Vendor Benchmarks Are Not ComparableWhy Anthropic, OpenAI and Google benchmark figures cannot be compared directly: differences in measurement and how to interpret the results.
How to decide when to change the AI model of a system in productionWhen to Change Your AI ModelHow to figure out if it's worth changing your AI model, when it's best to stop and how to make the transition without breaking anything.
Comparison of AI models from Anthropic, OpenAI and GoogleClaude, GPT and Gemini in comparisonCompare Claude, GPT and Gemini models: official prices, context windows, knowledge cutoffs and how to interpret comparisons across providers.
Comparison of Claude Fable 5.1 and Claude Sonnet 5 on price and capacityClaude Fable 5.1 vs Sonnet 5Claude Fable 5.1 vs Sonnet 5: a fivefold price difference with the same context window. When the extra cost is justified.
Comparison of context windows of major AI modelsContext Window: How Much It Really MattersWhat 1 million context tokens means, when you need them and why a large context window does not guarantee good answers.
What is Frontier-Bench and how does it measure the capabilities of AI agentsFrontier-Bench: what it is and what it measuresWhat Frontier-Bench is: the agentic task benchmark succeeding Terminal-Bench, and why it measures the agent and model as a pair.
Comparison of Gemini 3.6 Flash and Gemini 3.5 FlashGemini 3.6 Flash vs 3.5 FlashGemini 3.6 Flash vs 3.5 Flash: the same input price with cheaper output in the new version. When switching makes sense.
Comparison of GPT-5.6 Sol and GPT-5.6 Terra variantsGPT-5.6 Sol vs Terra: which makes sense?GPT-5.6 Sol vs GPT-5.6 Terra: official prices, differences and when the more expensive variant is justified.
Comparison between GPT-6 Astra and Claude Fable 5.1, same price listGPT-6 Astra vs Claude Fable 5.1Astra and Fable 5.1 cost the same. Compare official benchmarks to see where Astra wins and where Claude wins.
Comparison of GPT-6 Astra and GPT-5.6 Sol on official benchmarksGPT-6 Astra vs GPT-5.6 SolGPT-6 Astra vs GPT-5.6 Sol: a generational comparison of official benchmarks, prices and where improvements are visible.
Comparison of Claude Haiku 4.5 and Gemini Flash-Lite modelsClaude Haiku 4.5 vs Gemini Flash-LiteClaude Haiku 4.5 and Gemini Flash-Lite compared: official prices, service plans and when to choose each.
Comparison of Claude Haiku 4.5 and GPT-5.6 Luna in the economic beltClaude Haiku 4.5 vs GPT-5.6 LunaClaude Haiku 4.5 vs GPT-5.6 Luna: a fivefold input price difference, with very different context windows and knowledge cutoffs.
What AI models knowedge cutoff means and how it affects responsesKnowledge cutoff: what it means in practiceWhat a knowledge cutoff is, why models in the same family can have different dates and how to work with newer information.
How to Choose the Right AI Model for a Business AgentBest AI Model for AgentsChoose a model for an AI agent handling multi-step tasks: which benchmarks matter and what to measure in practice.
How to Choose the Right AI Model for Business Data AnalysisBest AI model for data analysisChoose an AI model for business data analysis, and understand why a plausible but incorrect analysis is the main risk.
How to Choose the Right AI Model for Customer SupportBest AI model for customer careChoose an AI model for customer support: why latency matters as much as quality and what the model must be able to do.
How to assess which AI model works best in ItalianBest AI Model for ItalianFind out which AI model performs best in Italian: why public benchmarks do not tell you and how to test it in an hour.
How to choose the right AI model based on your use caseThe best AI model depends on the use caseThere is no single best AI model. Choose according to your task, volume and latency, then compare results with a practical test before adopting it.
Comparison of low-cost AI models from Anthropic, OpenAI and GoogleCheap AI Models ComparedCompare low-cost AI models from Anthropic, OpenAI and Google: official prices, limitations and when they are sufficient.
How an SME chooses the right AI model for their processesWhich AI model suits an SMEHow an SME chooses the right AI model without technical expertise: three questions, a half-day test and mistakes to avoid.
How to Understand When an Economic AI Model Is EnoughWhen a low-cost AI model is enoughWhen a low-cost AI model is enough: a practical one-hour test to assess your task, compare results and choose without unnecessary spending.
Comparison of Claude Opus 5 and Claude Fable 5.1 on price and capacityClaude Opus 5 vs Fable 5.1Claude Opus 5 and Fable 5.1 compared: official prices, context, speed and practical criteria for choosing the right model.
Comparison of Claude Opus 5 and Gemini 3.1 Pro on price and featuresClaude Opus 5 vs Gemini 3.1 ProClaude Opus 5 vs Gemini 3.1 Pro: compare official rates, pricing tiers and service plans, with practical criteria for choosing the right model.
Comparison of Claude Opus 5 and GPT-6 Astra on price and featuresClaude Opus 5 vs GPT-6 AstraClaude Opus 5 vs GPT-6 Astra: official prices, context windows, knowledge cutoffs and how to choose between them.
Comparison of Claude Opus 5 and Claude Sonnet 5 on price and performanceClaude Opus 5 vs Sonnet 5Comparison between Claude Opus 5 and Sonnet 5: official prices, context, speed and when the cost difference is justified.
How the Reasoning Effort Level Changes the Results and Costs of an AI ModelReasoning effort: what changes in model comparisonsReasoning level changes the results and costs of the same AI model. Why a comparison that omits it is incomplete.
How SWE-bench and coding benchmarks work for AI modelsSWE-bench and coding benchmarksWhat SWE-bench is, how coding benchmarks work and what they reveal about a model's ability to write code.

The other categories