French AI company Mistral AI releases open-source model Mistral Medium3.5, but market response is lukewarm

BlockBeats News, May 1, French AI startup Mistral AI released its new generation open-source model Mistral Medium 3.5 on April 29. The model features a 128-billion-parameter dense architecture, and it also launched the remote coding agent tool Mistral Vibe CLI, as well as the multi-step AI Agent functionality for Le Chat.

According to official data, Mistral Medium 3.5 scored 77.6% on the SWE-Bench Verified programming benchmark, and it supports unified reasoning and coding capabilities. However, because its API pricing is as high as $1.5 per million input tokens and $7.5 per output token—significantly higher than that of Chinese open-source models—it has drawn widespread questioning from the community.

At present, leading positions in open-source model rankings are mainly occupied by Chinese models such as Alibaba’s Qwen, Zhipu GLM, and Xiaomi’s MiMo-V2. Among them, Qwen 3.6 has only 27 billion parameters, but it has already reached 72.4% on SWE-Bench, and it is fully open-source and free.

However, some developers believe that Mistral’s greatest value right now lies in its European identity as “not American, not Chinese.” Since it meets GDPR requirements and Europe’s local deployment needs, its models have been used for private deployments by major European institutions, including HSBC. Market views suggest that Mistral’s competitiveness comes more from geopolitics and corporate compliance rather than pure model performance.