Model evidence

Good enough for which task?

Explore measured models and available catalog listings. Quality comes from published task evidence; models without task results are clearly marked.

1205models listed13measured models1205matching models

Models, published task evidence, and observed provider pricing
Model	Evidence	Cheapest passing	Median quality	Providers
minimax/minimax-m2.1-highspeed Unclassified	Not yet measured	—	—	TrustedRouter
minimax/minimax-m2.5 Unclassified	Not yet measured	—	—	OpenRouter TrustedRouter
minimax/minimax-m2.5-highspeed Unclassified	Not yet measured	—	—	TrustedRouter
minimax/minimax-m2.7 Unclassified	Not yet measured	—	—	OpenRouter TrustedRouter
minimax/minimax-m2.7-highspeed Unclassified	Not yet measured	—	—	TrustedRouter
minimax/minimax-m2.7-turbo Unclassified	Not yet measured	—	—	TrustedRouter
minimax/minimax-m2.7:free Unclassified	Not yet measured	—	—	OpenRouter
minimax/minimax-m25 Unclassified	Not yet measured	—	—	TrustedRouter
minimax/minimax-m27 Unclassified	Not yet measured	—	—	TrustedRouter
minimax/minimax-m3 Unclassified	Not yet measured	—	—	OpenRouter TrustedRouter
minimax/minimax-m3-preview Unclassified	Not yet measured	—	—	TrustedRouter
minimax/minimax-m3:batch Unclassified	Not yet measured	—	—	OpenRouter
minimax/minimax-m3:free Unclassified	Not yet measured	—	—	OpenRouter
minimaxai/minimax-m1-80k Unclassified	Not yet measured	—	—	TrustedRouter
MiniMaxAI/MiniMax-M2.5 Unclassified	Not yet measured	—	—	TrustedRouter
MiniMaxAI/MiniMax-M3 Unclassified	Not yet measured	—	—	TrustedRouter
ministral-3-14b Unclassified	Not yet measured	—	—	Neurometric
mistral-7b Unclassified	Not yet measured	—	—	Neurometric
mistral-large-2402 Unclassified	Not yet measured	—	—	Neurometric
mistral-large-3-675b Unclassified	Not yet measured	—	—	Neurometric

Cheapest passing counts tasks where the model clears the task’s published quality floor (80% when none is recorded) and has the lowest observed workload cost. Median quality summarizes those published task results; it is not a general model ranking.

Model	Evidence	Cheapest passing	Median quality	Providers