Unverified50% confidenceBenchmarkExact time
After GRPO reinforcement learning, checkpoints using TailSFT achieved a final pass@1 improvement of up to 3.9 points.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/7/2026
First Seen
Valid until: 12/6/2026
Sources
Related Entities
Related Claims
UnverifiedAfter using TailSFT, the pass@16 metric on coding tasks during the SFT phase showed an absolute improvement of up to 16.8 points.71% similarUnverifiedGrok 4.6的智能指数比Grok 4.5高5分，比Grok 4.3高23分62% similarUnverifiedGrok 4.5 achieved 83.3% on Terminal Bench 2.1, surpassing Opus 4.8.61% similarUnverifiedGrok 4.5 produces 4.2 times fewer output tokens than Opus 4.8 on SWE Bench Pro tasks.61% similarUnverified根据官方在 WANDR 基准测试上的数据，Grok 4.6 位于性能与效率的帕累托前沿61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/873217API
curl https://kongchang.com/api/v1/knowledge/claims/873217MCP
get_claim(id=873217)