Best LLM for each job
The top three models for the jobs teams build most, from the Respan Index (cited public benchmarks) and each provider's list prices.
The top three models for the jobs teams build most, from the Respan Index (cited public benchmarks) and each provider's list prices.
Ranked by FrontierMath Tiers 1-3, original and unpublished math problems written and checked by mathematicians. Epoch AI runs every model on the same private set, so the scores compare like for like. Only models the Respan Index ranks are listed.
Price is the blended list price per 1M tokens, 3 parts input to 1 part output.
Each family's highest-ranked model for math, with its place in the ranking above.
The top two for math head to head, then the leading model families.
The ranking above uses FrontierMath. These public math benchmarks each have a full leaderboard, every score linked to its source.
GPT-6 Astra from OpenAI ranks first (math score 93.7%), followed by GPT-6.1 Sol (93.7%) and Claude Opus 5.5 (91.2%). Ranked by FrontierMath Tiers 1-3, original and unpublished math problems written and checked by mathematicians. Epoch AI runs every model on the same private set, so the scores compare like for like. Only models the Respan Index ranks are listed.
GPT-5.6 Luna scores 82.1%, 88% of the top score at 44x lower blended list price ($0.45 per 1M tokens). It is the cheapest model within 15% of the top score.
The best Claude model here is Claude Opus 5.5 at #3 (91.2%); the best GPT model is GPT-6 Astra at #1 (93.7%).
Ranked by FrontierMath Tiers 1-3, original and unpublished math problems written and checked by mathematicians. Epoch AI runs every model on the same private set, so the scores compare like for like. Only models the Respan Index ranks are listed. The FrontierMath leaderboard lists every score with a link to its source and is updated as new runs come out.
Yes. The Respan gateway serves models from every major provider behind one OpenAI-compatible API, with fallbacks when a provider fails.