1sec.ai
Back to feed
models1d ago

GLM-5.2 just dropped open weights and it already looks weirdly strong for coding

rr/LocalLLaMAscore 0.27

GLM-5.2, a text-only open-weights LLM, was released with a 1M context window and MIT license. Early results show it performing well in coding tasks, near the top of arenas. Its open nature allows for local deployment and testing on real-world repositories. You can download and test GLM-5.2 locally, which may be attractive for builders seeking an alternative to API-only models.

Key takeaways

  • 1M context window, open weights, and MIT license.
  • Performs well in early coding task benchmarks.
  • Allows for local deployment and testing on real-world code.
models1d ago

GLM-5.2 just dropped open weights and it already looks weirdly strong for coding

GLM-5.2, a text-only open-weights LLM, was released with a 1M context window and MIT license. Early results show it performing well in coding tasks, near the top of arenas. Its open nature allows for local deployment and testing on real-world repositories. You can download and test GLM-5.2 locally, which may be attractive for builders seeking an alternative to API-only models.

Key takeaways

  • 1M context window, open weights, and MIT license.
  • Performs well in early coding task benchmarks.
  • Allows for local deployment and testing on real-world code.