English(EN)⚖️ Qwen3.8 Max trails GPT-5.6 Sol by just 2.8 points, yet costs half per 1M output tokens. Open-weight is closing the gap without breaking the bank. https:// ol
AI 模型价格暴跌，开放权重模型日益实惠 · 跟踪 10 个来源
作者PulseAugur 编辑部·[17 个来源]·
在过去 30 天里，几款 AI 模型的输出 token 成本大幅降低，许多开放权重模型变得越来越实惠。腾讯发布了多款 Hy-MT2 模型（1.8B、7B、30B-A3B），价格具有竞争力，并提供 8.2k 的上下文窗口。Mistral Small 3.2 24B 和 Ling 3.0 Flash 的价格也大幅下跌，使其成为市面上最便宜的可用模型之一。就连 Gemini 3.6 Flash 等主要模型也将其输出定价减半，这表明 AI 模型市场正迅速朝着更低的成本趋势发展。
AI
影响
通过大幅降低开发人员和企业的运营成本，加速了 AI 模型的采用。
排序理由
包括开放权重模型在内的多款 AI 模型在短时间内输出 token 成本大幅降低，表明了市场趋势。
📉 Mistral Small 3.2 24B drops to $0.2/M output tokens — a 33% cut in 30 days. Open weights keep getting more affordable. See the latest pricing. O https:// olud.ai/model/mistral-small-3- 2-24b.html # AI # LLM # Pricing
Tencent just dropped Hy-MT2-1.8B with open weights. At $0.04/M in, it’s one of the cheapest small models with 8.2k context. See how it compares on the leaderboard. https:// olud.ai/latest.html # AI # LLM # OpenSource
⚖️ Qwen3.8 Max trails GPT-5.6 Sol by just 2.8 points, yet costs half per 1M output tokens. Open-weight is closing the gap without breaking the bank. https:// olud.ai/leaderboard.html # OpenSource # AI # LLM
🆕 New open-weight model just dropped: Hy-MT2-7B (Tencent) Context: 8.2k tokens · $0.07 in / $0.3 out per 1M · open weights All the latest models, tracked hourly: https:// olud.ai/latest.html # AI # LLM # OpenSource
⚡ New open model to know: Hy-MT2-30B-A3B (Tencent) Context: 8.2k tokens · $0.07 in / $0.3 out per 1M · open weights All the latest models, tracked hourly: https:// olud.ai/latest.html # AI # LLM # OpenSource
⚡ New open model to know: Hy-MT2-1.8B (Tencent) Context: 8.2k tokens · $0.04 in / $0.18 out per 1M · open weights All the latest models, tracked hourly: https:// olud.ai/latest.html # AI # LLM # OpenSource
📉 Gemini 3.6 Flash (batch) dropped to $1.88 per 1M output tokens — half of its price 30 days ago. That’s the kind of move that reshapes batch-cost math. https:// olud.ai/model/gemini-3-6-flash -batch.html # AI # LLM # Pricing
Gemini 3.6 Flash output pricing just halved in 30 days – from $7.50 to $3.75 per million tokens. That speed of drop is rare for a major model. https:// olud.ai/model/gemini-3-6-flash .html # AI # LLM # Pricing
Ling 3.0 Flash dropped 73% in 30 days to $0.06 per 1M output tokens – open weights. That price makes it one of the cheapest capable models, no subscriptions needed. https:// olud.ai/model/ling-3-0-flash.h tml # AI # LLM # Pricing
Qwen Plus 0728 (thinking) dropped from $1.2 to $0.78 per 1M output tokens (-35% in 30 days), and the weights are open. Open weights at that price is the rare part. See how it stacks up on the page that follows. https:// olud.ai/pricing.html # AI # LLM # Pricing
Hy3 dropped its output token price 38% in 30 days—now $0.33/M tokens with open weights. A rare move that makes a capable model far more accessible. https:// olud.ai/model/hy3.html # AI # LLM # Pricing
📉 Ling-3.0-flash: $0.22→$0.06 per 1M output tokens (-73% in 30 days). Open weights. That steep a drop signals fierce competition—the full pricing history is on olud.ai. https:// olud.ai/model/ling-3-0-flash.h tml # AI # LLM # Pricing
Qwen3 14B dropped 74% in 30 days — now $0.24 per 1M output tokens. Open weights at that price is hard to ignore. See the full trend. https:// olud.ai/model/qwen3-14b.html # AI # LLM # Pricing
Qwen3.8 27B just landed with 1M-token context and open weights. At $0.4 in / $3 out per 1M, it's a pricing shocker for a model this size—compare that to what you'd expect. https:// olud.ai/latest.html # AI # LLM # OpenSource
