Developer & API · Task Shortlists
Model shortlists
API models by task, in the order we’d test them. These are editorial judgments built from official specs and Artificial Analysis benchmarks, not task-specific measurements. Choosing a consumer chat app instead? Start with the App Picker →
01 · 4 candidates
General work
Your everyday starting point.
#1 GPT-6.2 SolClaude Opus 5.5Claude Sonnet 5.5+1
View ranked shortlist 02 · 7 candidatesCoding
Match the model to the change.
#1 Claude Sonnet 5.5Gemini 4 NeonClaude Opus 5.5+4
View ranked shortlist 03 · 4 candidatesResearch & analysis
Go deep, and check the sources.
#1 Claude Fable 5.2GPT-6.2 AstraGemini 4 Carbon+1
View ranked shortlist 04 · 4 candidatesWriting
Make it sound like you.
#1 Claude Sonnet 5.5Claude Opus 5.5GPT-6.1 Terra+1
View ranked shortlist 05 · 4 candidatesDocuments & office
The app matters as much as the model.
#1 GPT-6.2 AstraClaude Opus 5.5Gemini 4 Neon+1
View ranked shortlist 06 · 5 candidatesAgents & automation
Start with a cheaper model and move up when it fails.
#1 GPT-6.1 TerraClaude Haiku 5.5GPT-6.2 Sol+2
View ranked shortlist 07 · 4 candidatesHigh-volume pipelines
Cheap and fast, at scale.
#1 Claude Haiku 5.5GPT-6 LunaGemini 4 Hydrogen+1
View ranked shortlist 08 · 3 candidatesAudio & video input
Models that read audio and video.
#1 Gemini 4 NeonGemini 4 HydrogenGemini 4 Carbon
View ranked shortlist