Model comparisons
Which model, and what it costs
to finish the job.
Every pair I have run, indexed by model. Published price, the independent score, and the number that decides it in practice: what one finished task cost, not what a million tokens list for.
30 comparisons across 27 models. Each one is rewritten when the numbers move, not left to rot.
01 Nov 2026 · 4 min read
Claude Haiku 5.5 scores one point above Grok 4.7 at a fifteenth of the cost per task
01 Nov 2026 · 5 min readGemini 4 Carbon scores one point above Claude Opus 5.5 and costs 23% less per task, if you can get it
01 Oct 2026 · 7 min readClaude Sonnet 5.5 scores 18 points more than Sonnet 5 and the list price did not move
Pick the model you run
Models you can buy today. Each one lists what I have measured it against.
Claude Fable 5.1
Anthropic · 4 comparisonsClaude Opus 5.5
Anthropic · 4 comparisonsClaude Code
Anthropic · 1 comparisonClaude Haiku 5.5
Anthropic · 1 comparisonClaude Mythos 5.1
Anthropic · 1 comparisonClaude Sonnet 5.5
Anthropic · 1 comparisonGPT-6 Astra
OpenAI · 5 comparisonsCodex
OpenAI · 1 comparisonGPT-6 Luna
OpenAI · 1 comparisonGPT-6.1 Sol
OpenAI · 1 comparisonGemini 3.8 Flash
Google · 2 comparisonsGemini 4 Carbon
Google · 1 comparisonMuse Spark 1.3
Meta · 1 comparisonGrok 4.7
xAI · 1 comparisonDeepSeek V4.1 Flash
DeepSeek · 1 comparisonModels that have been replaced
Still here, because people run what they already bought. Each article says what took its place.
Claude Sonnet 5
GPT-5.6 Luna
Claude Haiku 4.5
Claude Sonnet 4.6
GPT-5.6 Terra
Claude Mythos 5
Claude Opus 4.8
Grok 4.6
In the order I ran them
Newest first, with the verdict on one line.
PairWhat it comes down toDate
Claude Haiku 5.5vsGrok 4.7one point up, 15x cheaper01 Nov 20264 min readGemini 4 CarbonvsClaude Opus 5.5first place, no API01 Nov 20265 min readClaude Sonnet 5.5vsClaude Sonnet 5same price, 18 points01 Oct 20267 min readGPT-6.1 SolvsGPT-6 Solfour points up, cache cut in half01 Oct 20266 min readGPT-6 AstravsClaude Opus 5.5better and dearer per task28 Sept 20268 min readClaude Opus 5.5vsOpus 5cheaper per token, dearer per task23 Sept 20269 min readGPT-6 LunavsClaude Haiku 4.51/10 price, 2x the score23 Sept 20267 min readGPT-6 SolvsClaude Opus 5.5the index and the invoice disagree23 Sept 20268 min readGPT-6 SolvsGPT-6 Astraties on code, loses the terminal23 Sept 20267 min readClaude CodevsCodexcoding agent benchmark and cost16 Sept 20267 min readThe systems behind these numbers are the actual work.
Model choice is one decision inside a revenue system. The rest of what I build and run is in the field notes.
Read the field notes →