Best AI for everyday questions (September 8, 2026)
September reopened a gap August had closed: Claude Fable 5.1 leads by more than two points, the widest lead since spring. Behind it the field is still packed - Opus 5, Muse Spark 1.3, GPT-6 Astra and Grok 4.6 all sit within two points of each other, and two Chinese labs remain in the top eight. Perplexity remains the specialist you add for anything that needs sources.
Ranking · high confidence
(Anthropic) · open ↗ · The new Opus 5 (July 24) tops the AA snapshot at 60.7 with Fable 5 right behind (59.9): Anthropic holds the top two spots. GPT-5.6 Sol follows (58.9) and stays the value pick for agentic work; Moonshot's Kimi K3 (57.1) remains the closest outside challenger.Free tier: Claude Sonnet 5 (usage limits vary by demand)- ChatGPT (OpenAI) · open ↗ · A hair behind on aggregate scores, and the most complete all-round app.Free tier: GPT-5.5 (limited messages, then a smaller model)
(Google) · open ↗ · Close third, with the most generous free tier of the three.Free tier: Gemini 3.6 Flash (most generous free tier, includes a monthly Deep Research allowance)
(Perplexity AI) · open ↗ · Not a generalist - but the first stop whenever the answer must be current and cited.Free tier: Sonar (unlimited basic cited searches + 5 Pro searches/day)
