亲眼看看 Spyingbee

OpenAI launched GPT-4.5 with native multimodal reasoning, while Anthropic responded with a 40% API price cut across all Claude models. Google DeepMind shipped Gemini 2.5 Pro with a 2M token context window. Meta AI released Llama 4 Scout (10M tokens, open-weight MoE), and Mistral AI undercut consumer pricing with Le Chat Pro at $14.99/mo.

This week marks the most aggressive pricing competition in the LLM market to date. Anthropic's 40% cut forces OpenAI to decide between margin and market share. Mistral's consumer pricing ($14.99 vs $20) targets cost-conscious developers. Meanwhile, Meta's open-weight Llama 4 Scout with 10M token context could commoditize long-context use cases that were previously enterprise-only. Google is playing the infrastructure card with EU data residency — a compliance moat that API-only competitors can't easily match.

Watch for OpenAI's pricing response within the next 2 weeks — their silence on Anthropic's cut is unusual. Mistral's Tokyo expansion signals serious enterprise ambitions beyond Europe. The real wildcard is Meta's Llama 4: if open-weight models close the gap with proprietary ones, it restructures the entire competitive landscape around fine-tuning services rather than base model access.