Unverified50% confidenceBenchmarkExact time
In tests on 20 Maven projects with deliberate defects, Qwen2.5-Coder 7B successfully fixed 9/20 (45%, average latency 29 seconds), Qwen2.5-Coder 1.5B successfully fixed 4/20 (20%, average latency 12 seconds), and DeepSeek-Coder 6.7B successfully fixed 1/20 (5%, average latency 39 seconds).
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/15/2026
First Seen
Valid until: 10/29/2026
Sources
Related Entities
Related Claims
UnverifiedIn a 14-question test for fixing real bugs, DevStral and Qwen3 each fixed 12, with an average of no more than 97 seconds per question.70% similarUnverifiedExperience internalization fine-tuning improved Qwen3-8B-27B's one-shot score from 37.07 to 48.22, while reducing generated tokens from 6.41 million to 570,000 (an 11-fold reduction).66% similarUnverifiedAccording to the SWE-Bench Verified benchmark, Qwen 2.5 Coder 32B achieves a coding ability score of 61.765% similarUnverifiedQwen 2.5 Coder, with only 32B parameters, matched the larger general-purpose model GPT-4o, indicating that code-specialized training has a punch-above-its-weight effect on complex array shape and broadcasting errors.65% similarUnverifiedQwen3 27B achieved a 100% pass rate in the 256K long-context memory retrieval test (needle in a haystack), with all 15 runs being accurate.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1036233API
curl https://kongchang.com/api/v1/knowledge/claims/1036233MCP
get_claim(id=1036233)