DeepSeek V4.1 Flash API is free.
And it just beat Claude Opus at coding benchmarks. The numbers: 74.2 on SWE-bench. Claude Opus scored 74. GPT 5.6 scored 73. 88.1 on CyberGym. Ahead of both. It's a 552B parameter model that only wakes up 8B at a time. That's why it's fast. 1
DeepSeek just retired its biggest model.
A smaller free one beat it. V4.1 Flash outperformed V4 Pro on speed, performance, and runtime. Now Pro requests route to Flash automatically. Here's how it wins: It's a 552 billion parameter model. But it only wakes up 8 billion at a
Rough Seas: Ship Buoyancy Simulation
Live Link: Details: Model and Harness: deepseek_ai-v4-Flash in NousResearch Hermes Agent Cost - $0.51 USD (DeepSeek Official API) API requests - 296 Tokens - 73,457,748 Time - 2 h 05 min
How to run GLM-5.3 flash, deepseek V4 flash and other 321B frontier model on your own machine 😳
With colibri repo. 34,227 stars. pure C, no GPU, no api key, no per token bill what $0 gets you: -GLM-5.3 Flash, 321B params, ~195 GB converted -vision included -DeepSeek V4 Flash,
You have reached the end of the archive
All of deepseek