Unverified50% confidenceBenchmarkExact time
Using only BM25/grep and MiniMax M3, without a compiled index or vector index, one run achieved 85% accuracy on BrowseComp-Plus.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/2/2026
First Seen
Valid until: 1/31/2027
Sources
Related Entities
Related Claims
UnverifiedUnder an 18k character budget, anchored hybrid retrieval achieved 36.18% accuracy, BM25 achieved 35.98%, while the best direct planner-guided method reached only 34.19%.70% similarUnverifiedThe ratio of activated parameters to total parameters for MiniMax M3, 23B/428B, is approximately 5.4%.65% similarUnverifiedQuantization methods such as GPTQ, AWQ, and SqueezeLLM can still maintain over 95% of the original model's performance at 4-bit precision.65% similarUnverifiedLlama 3.1 405B's score on MMLU is already close to the level of GPT-4o, and Qwen 2.5 72B has surpassed GPT-4 Turbo on certain metrics in math and code generation benchmarks.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1125087API
curl https://kongchang.com/api/v1/knowledge/claims/1125087MCP
get_claim(id=1125087)