Kimi K2.6 has some real value for the price, costing $0.69 to reach 72.5%, which is one-third cheaper than GLM. Domestic models are starting to compete seriously.

GLM-4.67%
View Original
CoinNetwork
Ramp releases SWE-Bench private benchmark: Claude Fable 5 wins with a 87.5% win rate
Ramp releases a private testing benchmark SWE-Bench, containing 80 backend development tasks from real production environments, aimed at solving issues of public evaluation data leakage and metric saturation.
14 models were evaluated, showing Anthropic Claude Fable 5 leading at 87.5%, with Claude Opus 4.7 and GPT-5.5 tied at 83.75%.
Domestic Kimi K2.6 and GLM 5.1 have solution rates of 72.5% and 71.25%, respectively, but Kimi K2.6 costs $0.69 on average, about 34% cheaper than GLM 5.1.
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pinned