AINews Portal

⚖️ Compare models

Side-by-side specs, arena Elo, and cheapest prices for up to 4 models. Best value in each row is highlighted.

SpecR1 Distill Llama 70B✕R1 0528✕
Organizationdeepseekdeepseek
Use casereasoningreasoning
Capabilitiesreasoning, chatreasoning, chat
AccessOpen weights (unrestricted)MIT
Text Arena Elo—1428
Cheapest input $/M$0.20$0.20
Cheapest output $/M$0.38$0.25
Context window128K164K
Parameters70B—
HF downloads41.8K172.4K
7d news mentions——
← All models