τ²-bench — cost vs performance

Best current row per canonical model in the group “pass^1 · evaluator=Artificial Analysis” (323 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.

← Leaderboard

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

open / restricted weights closed Pareto frontier (8 models)

Frontier models 8

Pareto frontier
Model	pass^1 on τ²-bench	Rank	cheapest current output price (USD / 1M tokens)	Provider	Trust
Z.ai GLM 5.2Z.ai (Zhipu AI) · Open weights	99.1%	1	$2	Z.ai API	Independent
GLM 4.7 FlashZ.ai (Zhipu AI) · Open weights	98.8%	3	$0.40	Z.ai API	Independent
Step 3.5 FlashStepFun · Open weights	94.4%	28	$0.30	OpenRouter	Independent
MiMo-V2.5Xiaomi · Open weights	90.6%	54	$0.28	OpenRouter	Independent
Qwen3.5-9BQwen · Open weights	86.8%	72	$0.15	OpenRouter	Independent
gpt-oss-20bOpenAI · Open weights	60.2%	131	$0.13	OpenRouter	Independent
Ministral 3 3BMistral AI · Open weights	24.9%	223	$0.10	Mistral AI La Plateforme	Independent
Mistral Small 3Mistral AI · Open weights	19.6%	253	$0.08	Mistral AI La Plateforme	Independent

Methodology. Points are the best current row per canonical model in comparability group 'pass^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.