Skip to main content
idapt
AccueilModèles IATarifs
Sign inStart free
All terms

Learn

What is MATH-500?

Last reviewed July 16, 2026

MATH-500 is a benchmark of 500 competition mathematics problems sampled from the larger MATH dataset, spanning subjects like algebra, geometry, number theory, and precalculus across graded difficulty levels. Each problem has a checkable final answer, so scoring is objective. It sits in the middle of the math benchmark ladder: harder than grade-school word problems, well below research mathematics.

Where it sits on the ladder

The math ladder runs roughly: grade-school arithmetic (saturated long ago), MATH-500 competition problems, AIME-level contest exams, then research-grade sets like FrontierMath. Frontier reasoning models now score very high on MATH-500, so it differentiates mid-tier and small models best, while the top of the field is measured on the harder rungs.

Difficulty banding

Problems carry difficulty levels from the original MATH dataset, so evaluators can report performance per band. A model can be reliable on the easy bands and collapse at the hardest; the aggregate score hides that shape. For workloads with real math in them, hard-band behavior matters more than the average.

Top models on MATH-500

Full leaderboard
  1. 1GPT 598.1%
  2. 2GPT 5 Mini97.8%
  3. 3o4-mini (High)97.8%
  4. 4o397.8%
  5. 5Qwen3 Max Thinking97.1%
  6. 6DeepSeek R1 (0528)96.6%
  7. 7o3-mini (High)96.5%
  8. 8Gemini 2.5 Pro (05-06)95.9%
  9. 9GPT 5 Nano95.2%
  10. 10o194.7%

Peak scores from the compiled benchmark data; reasoning models use their highest effort tier.

In idapt

idapt's math rankings include a MATH-500 lane beside AIME and FrontierMath, with prices next to scores, so you can see which cheaper models already clear this bar.

MATH-500 leaderboardMath rankings

Frequently asked

Related terms

AIMEFrontierMath

Multi-Model Chat

GPT, Claude, Gemini, Grok, Llama: pick mid-conversation, cheapest-first.

Code Execution

Sandboxed Python, Node, and shell: agents and you, side-by-side.

See it in practice

The fastest way to understand it is to use it. Start free, no credit card.

Try idapt free
  • Accueil
  • Tarifs
  • Modèles IA
  • Modèles d'image
  • Modèles de voix
  • Modèles vidéo
  • Classements
  • New models
  • Model status
  • Multi-Model Chat
  • Voice Mode
  • Voice HUD
  • Web Search
  • Image Generation
  • Video Generation
  • Audio Generation
  • Transcription
  • Drive
  • Secrets
  • Sharing
  • Workspaces
  • Tasks
  • Memory
  • Agents
  • Subagents
  • Automations
  • Skills
  • Code Execution
  • Computers
  • Computer Use
  • Cloud Computers
  • Local AI
  • AI Gateway
  • API & SDK
  • CLI
  • MCP
  • Tunnels
  • All features →
  • Blog
  • Cas d'usage
  • Comparisons
  • Best of
  • Learn
  • Free tools
  • Prompts
  • Templates
  • Changelog
  • Centre d'aide
  • FAQ
  • Confidentialité
  • Comparer tous les modèles
  • Support
  • Développeurs
  • Quickstarts
  • Référence API
  • Tarifs API
  • CLI
  • MCP
  • Downloads
  • Desktop
  • Badges and embeds
© idapt[email protected]ConditionsPolitique de confidentialitéMentions légalesSignaler un contenu
X (Twitter)