Unverified75% confidenceFactTime unknown
使用Gemma 4 2B版本本地运行Codex时，处理速度大约每秒30到40个Token
1
Sources
75%
Confidence
Long-term
Relevance
6/3/2026
First Seen
Sources
Ollama+Gemma 4本地运行Codex：零成本AI编程完整指南
bilibili薛定猫AI
Related Entities
Related Claims
Unverified在24GB显存的显卡上运行Qwen3.6 27B模型，未经优化可达每秒约40个Token的生成速度，优化后可达50-60 Token/s70% similarUnverifiedClaude Haiku 4.5的处理速度可达每秒100-200+ token68% similarUnverified在配备96GB显存的M2 Max上以1-bit精度运行时，简单任务生成速度稳定在约28-30 tokens/秒68% similarUnverifiedQwen3.6 35B模型借助MTP技术在oMLX上达到了86.7 tokens/s的生成速度，prompt处理速度为1735 tokens/s67% similarUnverifiedRTX 5090上Qwen3.6 27B模型Q4量化不开MTP解码速度约63 token/s，开启MTP后跳到105 token/s67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/38842API
curl https://kongchang.com/api/v1/knowledge/claims/38842MCP
get_claim(id=38842)