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 the Respan Index agents score, which combines agent benchmarks such as OSWorld-Verified, tau2-bench, Toolathlon and APEX-Agents. Only models with published agent results are ranked.
Price is the blended list price per 1M tokens, 3 parts input to 1 part output.
Each family's highest-ranked model for agents & tool use, with its place in the ranking above.
The top two for agents & tool use head to head, then the leading model families.
17 public benchmarks feed it. The core set, run independently on many models, anchors the scale; the rest add coverage. Each links to its full leaderboard where there is one.
Claude Fable 5.2 from Anthropic ranks first (agents score 85.3), followed by Gemini 4 Neon (84.5) and GPT-6.2 Astra (83.6). Ranked by the Respan Index agents score, which combines agent benchmarks such as OSWorld-Verified, tau2-bench, Toolathlon and APEX-Agents. Only models with published agent results are ranked.
DeepSeek-V4.1-Flash-1017 scores 76.3, 89% of the top score at 159x lower blended list price ($0.1 per 1M tokens). It is the cheapest model within 15% of the top score.
The best Claude model here is Claude Fable 5.2 at #1 (85.3); the best GPT model is GPT-6.2 Astra at #3 (83.6).
Ranked by the Respan Index agents score, which combines agent benchmarks such as OSWorld-Verified, tau2-bench, Toolathlon and APEX-Agents. Only models with published agent results are ranked. The Respan Index combines published benchmark results, each linked to its source, and is updated as new results come out.
Yes. The Respan gateway serves models from every major provider behind one OpenAI-compatible API, with fallbacks when a provider fails.