Unverified50% confidenceFactExact time
The new version of Mistral Large 4 used 30% fewer tokens in its output.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/31/2026
First Seen
Valid until: 1/29/2027
Sources
Mistral Large 4 Open Weights: Trillion-Parameter MoE, Sharp Gains in Terminal Coding
bilibiliSmallSmall_10/30/2026
Related Entities
Related Claims
UnverifiedMistral Large 4's AA score is 38, a significant improvement over the previous generation Mistral Medium 3.5's score of 14.77% similarUnverifiedMistral Large 4 uses an average of 47% fewer tokens per task on equivalent tasks.73% similarUnverifiedOn the "can the change be merged" test, Mistral Large 4 scored 19.7%, only 0.4 percentage points above second place.73% similarUnverifiedFrance's Mistral has released Mistral Large 4, scoring 38 points in a third-party intelligence index evaluation.72% similarUnverifiedMistral Large 4 scored 28% on TerminalBench, surpassing Kimi K3 and QN 3.8 Max, both of which are larger in model size than Mistral Large 4.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1119757API
curl https://kongchang.com/api/v1/knowledge/claims/1119757MCP
get_claim(id=1119757)