Terminal-Bench — cost vs performance

Best current row per canonical model in the group “accuracy · variant=v2.1 · evaluator=Artificial Analysis” (183 models) against total parameters. The dashed line is the Pareto frontier: no model is both better and cheaper than a point on it.

← Leaderboard

XOutput price Input price Parameters Context Memory (est.)Scalelog linearBubblecontext params none

open / restricted weights closed Pareto frontier (7 models)bubble = context window

Frontier models 7

Pareto frontier
Model	accuracy on Terminal-Bench	Rank	total parameters	Provider	Trust
Kimi K3Moonshot AI · Open weights	85.0%	11	2.78T	—	Independent
GLM 5.3 FlashZ.ai (Zhipu AI) · Open weights	84.3%	14	321.3B	—	Independent
Qwen3.8 27BQwen · Open weights	79.8%	25	27.8B	—	Independent
Gemma 4 26B A4BGoogle · Open weights	39.0%	96	25.8B	—	Independent
Qwen3.5-9BQwen · Open weights	29.2%	110	9.65B	—	Independent
Qwen3.5-4BQwen · Open weights	25.8%	118	4.66B	—	Independent
LFM2.5-2.6B (free)Liquid AI · Open weights	4.49%	157	2.7B	—	Independent

Methodology. Points are the best current row per canonical model in comparability group 'accuracy · variant=v2.1 · 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.