Skip to content
AI Atlas

Terminal-Bench — cost vs performance

Best current row per canonical model in the group “accuracy · variant=hard · evaluator=Artificial Analysis” (317 models) against cheapest current output price (USD / 1M tokens). The dashed line is the Pareto frontier: no model is both better and cheaper than a point on it.

XOutput priceInput priceParametersContextMemory (est.)ScaleloglinearBubblecontextparamsnone
Group

open / restricted weights closed Pareto frontier (12 models)bubble = context window

Frontier models 12

Pareto frontier
Modelaccuracy on Terminal-BenchRankcheapest current output price (USD / 1M tokens)ProviderTrust
gpt-5.6-solOpenAI · Closed65.9%1$5OpenAI APIIndependent
gpt-5.4-miniOpenAI · Closed52.3%11$2.25OpenAI APIIndependent
Z.ai GLM 5.2Z.ai (Zhipu AI) · Open weights50.8%12$2OpenRouterIndependent
KAT-Coder-Pro V2Kwaipilot · Closed49.2%14$1.2OpenRouterIndependent
MiMo-V2.5-ProXiaomi · Open weights43.2%27$0.87OpenRouterIndependent
gpt-5.4-nanoOpenAI · Closed42.4%30$0.625OpenAI APIIndependent
MiMo-V2.5Xiaomi · Open weights41.7%33$0.28OpenRouterIndependent
Gemma 4 26B A4BGoogle · Open weights25%100$0.22Google Gemini APIIndependent
Qwen3.5-9BQwen · Open weights24.2%105$0.15OpenRouterIndependent
gpt-oss-20bOpenAI · Open weights10.6%164$0.13OpenRouterIndependent
Granite 4.0 MicroIBM · Open weights1.52%256$0.112OpenRouterIndependent
Gemma 3 4BGoogle · Restricted weights0.76%265$0.10Google Gemini APIIndependent

Methodology. Points are the best current row per canonical model in comparability group 'accuracy · variant=hard · evaluator=Artificial Analysis'. Price = cheapest current offer across providers (the provider shown). Pareto frontier maximises the score and minimises x; exact ties are all kept. memory_estimate is an estimate (see /methodology). Only the selected comparability group is plotted; points under other configurations are not mixed in. Price = cheapest current offer across providers at the time of the last crawl. Nothing is estimated except where marked.