Chat with the newest AI models and work with them on your hardest problems

July 16, 2026 · GPT-5.6 Sol is now available in chat. Premium models GPT-5.6 Sol Pro and Claude Fable 5 are now available on request.

May 27, 2026 · Benchmarks in Leipzig is public — 100 research-level questions from 49 contributors (results, arXiv).

Every user gets free monthly credit. Prepaid credit is available individually or as a shared pool for research groups and teams.

See My Chat or contact us for shared pools at contact@sciencebench.ai.

Newest Available Frontier Models:

GPT-5.6 Sol ∅ per turn: 0.91cp Claude Opus 4.8 ∅ per turn: 0.94cp Grok-4.3 ∅ per turn: 0.33cp Gemini 3.1 Pro ∅ per turn: 0.45cp

Premium Models (only available for grant users):

Claude Fable 5 ∅ per turn: ~3.5cp GPT-5.6 Sol Pro ∅ per turn: ~10cp

313

researchers signed up

325

benchmark problems by 102 users

3455

chat turns by 141 users

See how the newest models perform on research-level mathematics.

Browse research-level problems contributed by the community.