Agentic model routing

Find the right model for agent work

Match a workload to models using Agentic Index, task-level cost, response time, benchmark signals, and context requirements. The shortlist is built from the same live Artificial Analysis data used across WhatLLM.

Agentic fit lab

Choose the model that fits the work

Ranking changes as task shape, autonomy, context, and economics change.

Task

Autonomy needed

Context load

Optimize for

Frontier map

Agentic Index vs task cost

23 high-fit models

37 with cost/task

20324355677890$0.010$0.030$0.100$0.300$1.00$3.00Agentic IndexCost per taskGPT-5.5 (high) · Agentic 72 · $0.668 per taskClaude Opus 4.8 (Adaptive Reasoning, Max Effort) · Agentic 77.8 · $1.78 per taskGPT-5.5 (xhigh) · Agentic 74.1 · $0.993 per taskClaude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) · Agentic 80.6 · $3.25 per taskClaude Opus 4.7 (Adaptive Reasoning, Max Effort) · Agentic 71.3 · $1.97 per taskQwen3.7 Max · Agentic 66.6 · $0.456 per taskGPT-5.4 (xhigh) · Agentic 68 · $1.03 per taskGemini 3.5 Flash (high) · Agentic 70.3 · $0.614 per taskDeepSeek V4 Pro (Reasoning, Max Effort) · Agentic 67.2 · $0.056 per taskGemini 3.1 Pro Preview · Agentic 59.1 · $0.305 per taskMiniMax-M3 · Agentic 68.6 · $0.182 per taskMiMo-V2.5-Pro · Agentic 67.4 · $0.062 per taskGLM-5.1 (Reasoning) · Agentic 67.1 · $0.270 per taskGPT-5.4 mini (xhigh) · Agentic 58.9 · $0.479 per taskQwen3.7 Plus · Agentic 65.1 · $0.057 per taskGrok 4.3 (high) · Agentic 65.9 · $0.145 per taskClaude Sonnet 4.6 (Adaptive Reasoning, Max Effort) · Agentic 63 · $1.14 per taskKimi K2.6 · Agentic 66 · $0.294 per taskMiniMax-M2.7 · Agentic 61.5 · $0.088 per taskDeepSeek V4 Flash (Reasoning, Max Effort) · Agentic 61.3 · $0.043 per taskQwen3.6 Plus · Agentic 61.7 · $0.161 per taskNemotron 3 Ultra 550B A55B (Reasoning) · Agentic 57.1 · $0.244 per taskStep 3.7 Flash · Agentic 59.5 · $0.131 per taskQwen3.6 27B (Reasoning) · Agentic 62.9 · $0.257 per taskQwen3.5 397B A17B (Reasoning) · Agentic 55.8 · $0.333 per taskQwen3.5 122B A10B (Reasoning) · Agentic 53 · $0.240 per taskGPT-5.4 nano (xhigh) · Agentic 47.6 · $0.153 per taskMistral Medium 3.5 · Agentic 53.2 · $0.491 per taskRing-2.6-1T · Agentic 51.5 · $0.344 per taskGrok 4.3 (Non-reasoning) · Agentic 48.8 · $0.529 per taskNova 2.0 Pro Preview (medium) · Agentic 47 · $0.173 per taskClaude 4.5 Haiku (Reasoning) · Agentic 40.2 · $0.207 per taskNVIDIA Nemotron 3 Super 120B A12B (Reasoning) · Agentic 40.2 · $0.218 per taskgpt-oss-120b (high) · Agentic 37.9 · $0.061 per taskGemma 4 26B A4B (Reasoning) · Agentic 32.1 · $0.032 per taskGemini 3.1 Flash-Lite · Agentic 25.7 · $0.043 per taskgpt-oss-20B (high) · Agentic 27.6 · $0.018 per taskGPT-5.5 (high)Claude Opus 4.8 (max)GPT-5.5 (xhigh)Claude Fable 5 (Max EfClaude Opus 4.7 (max)

GPT-5.5 (high)

OpenAI

Agentic

72

Task cost

$0.668

ModelFitAgenticTask costResponseContext
GPT-5.5 (high)

OpenAI · Proprietary

9972$0.6681.0m922K
Claude Opus 4.8 (max)

Anthropic · Proprietary

9977.8$1.7843.4s1.0M
GPT-5.5 (xhigh)

OpenAI · Proprietary

9774.1$0.9932.1m922K
Claude Fable 5 (Max Effort, Opus 4.8 Fallback)

Anthropic · Proprietary

9680.6$3.25N/A1.0M
Claude Opus 4.7 (max)

Anthropic · Proprietary

9271.3$1.9728.7s1.0M
GPT-5.5 (medium)

OpenAI · Proprietary

9269.4N/A17.2s922K
Qwen3.7 Max

Alibaba · Proprietary

9166.6$0.45619.0s1.0M
GPT-5.4 (xhigh)

OpenAI · Proprietary

9068$1.032.0m1.1M
Gemini 3.5 Flash (high)

Google · Proprietary

9070.3$0.61422.8s1.0M
DeepSeek V4 Pro (Reasoning, Max Effort)

DeepSeek · Open

8767.2$0.0561.1m1.0M

Why cost per task changes the decision

Agent loops compound price and latency

A small per-token price gap can become a large bill when an agent runs many turns, calls tools, and carries long state. Cost per Intelligence Index task gives a cleaner decision unit than token price alone.

The best model depends on the failure cost

Frontier models are worth it when mistakes are expensive. For routine automation, a cheaper high-fit model can preserve most of the capability while cutting task cost sharply.