Which AI should you use?

Start with what you want to get done. We'll match your task to the evidence, not to the biggest marketing budget.

In a hurry? Just tell me: two quick choices, one answer.

What do you need help with?

Here are the options worth considering.

We've ranked these using the tests that are relevant to your task. Each test has limits, so treat this as a useful starting point, not a guarantee.

In short

For homework and maths, our pick is GPT-6.1 Sol. GPT-6 Astra and Claude Opus 5.5 are too close to call against it. If you would rather use a model with open weights, Kimi K3 is the best of those we can see. We have verified prices for GPT-6.1 Sol and GPT-6 Astra only; for the rest we cannot say what they cost.

For Homework and maths, we looked at hard maths problems first.

Best-supported choice

GPT-6.1 Sol by OpenAI

Scored highest on hard maths problems among the models with enough evidence, and none of its other relevant results put it behind the leaders.

Worth knowing: In the middle of the listed models for writing: 1,425 on Arena text: creative writing, 54th of 77.

Price:
  • ChatGPT Plus: GBP 20.00 per month (price source, verified 29 Oct 2026)
  • ChatGPT Pro (from): GBP 89.00 per month (price source, verified 29 Oct 2026)
How we know: scores and sources
Test results for GPT-6.1 Sol relevant to homework and maths
TestResultRatingSetting usedSource
FrontierMath Tiers 1-3 (v2)93.7% share of problems solvedTop tiermaxSource for GPT-6.1 Sol FrontierMath Tiers 1-3 (v2) score (epoch)
OTIS Mock AIME 2024-2025100.0% share of problems solvedTop tiermaxSource for GPT-6.1 Sol OTIS Mock AIME 2024-2025 score (epoch)
Arena text: maths1,486 Arena scoreStrongMaxSource for GPT-6.1 Sol Arena text: maths score (lmarena)

See everything about GPT-6.1 Sol

Also worth a look

Good-enough alternative Close call with our top pick

GPT-6 Astra by OpenAI

Rated Top tier on hard maths problems.

Worth knowing: On maths votes it sits in the middle of the listed models.

Price:
  • ChatGPT Plus: GBP 20.00 per month (price source, verified 29 Oct 2026)
  • ChatGPT Pro (from): GBP 89.00 per month (price source, verified 29 Oct 2026)
How we know: scores and sources
Test results for GPT-6 Astra relevant to homework and maths
TestResultRatingSetting usedSource
FrontierMath Tiers 1-3 (v2)93.7% share of problems solvedTop tiermaxSource for GPT-6 Astra FrontierMath Tiers 1-3 (v2) score (epoch)
OTIS Mock AIME 2024-2025100.0% share of problems solvedTop tiermaxSource for GPT-6 Astra OTIS Mock AIME 2024-2025 score (epoch)
Arena text: maths1,468 Arena scoreMiddleMaxSource for GPT-6 Astra Arena text: maths score (lmarena)

Full profile for GPT-6 Astra

Good-enough alternative Close call with our top pick

Claude Opus 5.5 by Anthropic

Rated Top tier on hard maths problems.

Worth knowing: Its score is within the margin of error of the best-supported choice, so the order between them is not settled.

Price:

Price not independently verified

How we know: scores and sources
Test results for Claude Opus 5.5 relevant to homework and maths
TestResultRatingSetting usedSource
FrontierMath Tiers 1-3 (v2)91.2% share of problems solvedTop tiermaxSource for Claude Opus 5.5 FrontierMath Tiers 1-3 (v2) score (epoch)
OTIS Mock AIME 2024-2025100.0% share of problems solvedTop tiermaxSource for Claude Opus 5.5 OTIS Mock AIME 2024-2025 score (epoch)
Arena text: maths1,517 Arena scoreTop tierHighSource for Claude Opus 5.5 Arena text: maths score (lmarena)

Full profile for Claude Opus 5.5

Best open-weights option

Kimi K3 by Moonshot

Rated Strong on hard maths problems. Its weights are published under an open licence, so it can be offered by more than one provider. We have not verified any price or free access.

Worth knowing: Not enough data for fixing real software bugs, so part of the picture is missing.

Price:

Price not independently verified

How we know: scores and sources
Test results for Kimi K3 relevant to homework and maths
TestResultRatingSetting usedSource
FrontierMath Tiers 1-3 (v2)72.2% share of problems solvedStrongmaxSource for Kimi K3 FrontierMath Tiers 1-3 (v2) score (epoch)
OTIS Mock AIME 2024-202597.2% share of problems solvedStrongmaxSource for Kimi K3 OTIS Mock AIME 2024-2025 score (epoch)
Arena text: maths1,492 Arena scoreTop tierMaxSource for Kimi K3 Arena text: maths score (lmarena)

Full profile for Kimi K3

About cost. We can compare performance, but we don't yet have enough verified pricing data to tell you which option is cheapest. We never use API prices as a stand-in for consumer subscription prices, and we never assume a free tier has the same features, limits or model access as a paid plan. New to tokens and “per million”? Read how AI pricing works.

Other models with enough evidence (39)

These have a result for the main test but were not shown above. They rated Middle or Behind the leaders on it, or fell outside the top few. Listed in score order.

Models we could not rank for this task (34)

Missing evidence is not a low score. These models are left out of the ranking because the tests for this task have no result for them.

How this was decided: our recommendation rules. Results are relative to the 77 models listed here.

Your choice stays in your browser. Nothing you select here is sent anywhere.