Unverified50% confidenceBenchmarkExact time
Mistral Large 4 scored approximately 62% on the real-world software engineering benchmark DeepSwee.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/8/2026
First Seen
Valid until: 1/6/2027
Sources
Related Entities
Related Claims
UnverifiedMistral Large 4 narrowly outperformed GPT-6 Astra on visual grounding tasks with a score of 42% versus 41%.74% similarUnverifiedIn the Arena AI blind-test ELO ranking (coding and web development), Mistral Large 4 ranks 45th, with performance roughly equivalent to Opus 4.8, Sonnet 4.6, or GPT-5.6 level.71% similarUnverifiedMistral Large 4 solved 18 out of 19 capture-the-flag (CTF) challenges, and in a malware reverse-engineering test identified Cobalt Strike, extracted configurations and indicators of compromise, and generated an investigation report containing Yara rules within 12 minutes.69% similarUnverifiedOn the DeepSWE 1.1 coding benchmark, Mistral Large 4 Preview scored 62, ranking second, behind only Kimi K3 paired with Kimi Code CLI at 68.68% similarUnverifiedDeepSeek V4.1 claims to achieve 74 on SWE, but the blogger found that this was the best result cherry-picked from multiple tests, with open code testing at 65 and PI testing at 66.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/987788API
curl https://kongchang.com/api/v1/knowledge/claims/987788MCP
get_claim(id=987788)