Mistral Small 3.2 24B
Mistral Small 4Newer version
GPT-6.1 TerraCheaper, as goodGPT-6 LunaBetter
Qwen3.8 27BOpen weightsMistral Small 3.2 24BRankingin non-thinking mode?
Alternatives?
CheaperGPT-6.1 Terramedium×5.9-13%$0.095BetterGPT-6 Lunalow×3.1-96%$0.005Function calling only with none · Chat Completions+3OpenQwen3.8 27Bxhigh×4.8+950%$1.15Memory (4-bit): 14 GB → 17 GB+5
CheaperMiMo v2.6 FlashthinkingOpen×5.4-43%$0.062Memory (4-bit): 14 GB → 186 GB+4
CheaperClaude Haiku 5.5high×5.4-27%$0.079Remove sampling parameters+5GPT-6.1 Terramedium×5.9-13%$0.095GPT-6 Lunamax×5.5-38%$0.068Function calling only with none · Chat Completions+3MiMo v2.6 FlashthinkingOpen×5.4-43%$0.062Memory (4-bit): 14 GB → 186 GB+4Claude Haiku 5.5high×5.4-27%$0.079Remove sampling parameters+5GPT-5.6 Lunaxhigh×5-22%$0.085GPT-6 Lunalow×3.1-96%$0.005Function calling only with none · Chat Completions+3
Granite 4.2 3BthinkingOpen+30%-95%$0.006No image inputGPT-5.6 Lunalow×3-91%$0.01GPT OSS 20BhighOpen+29%-90%$0.011No image inputMistral Small 4thinkingOpen+62%-84%$0.017Memory (4-bit): 14 GB → 72 GBQwen3.8 27BxhighOpen×4.8+950%$1.15Memory (4-bit): 14 GB → 17 GB+5IK2 Horizon MoVA 36B A4BthinkingOpen×3.6+1,067%$1 / $4Memory (4-bit): 14 GB → 22 GB+1Qwen 3.5 27BthinkingOpen×3.3+450%$0.30 / $2.40Memory (4-bit): 14 GB → 17 GBAG9v3 39A5BOpen×3.1——Memory (4-bit): 14 GB → 23 GB+1Qwen3.6 27BthinkingOpen×3.1+469%$0.62Memory (4-bit): 14 GB → 17 GBMistral Large 4thinking×5.5+1,371%$1.61Infinite whitespace stream+4Mistral Medium 3.5highOpen×2+293%$0.43Memory (4-bit): 14 GB → 77 GBMistral Small 4thinkingOpen+62%-84%$0.017Memory (4-bit): 14 GB → 72 GBMistral Large 3non-thinkingOpen+33%-71%$0.031Memory (4-bit): not verifiedMagistral Medium 1thinking+30%——No image inputDevstral Mediumnon-thinkingOpen+29%—$0.40 / $2Memory (4-bit): not verified+1Magistral Small 1thinkingOpen+18%—$0.70 / $1.16No image inputMistral Small 3non-thinkingOpensame—$0.05 / $0.08Context: 33K → 32K+1All Mistral models →Run it yourself
estimateBenchmarks
all categoriesFAQ
How much does Mistral Small 3.2 24B cost?›
$0.10 per 1M input tokens and $0.30 per 1M output tokens; cached input $0.01 per 1M; cheapest via DeepInfra at $0.075 / $0.20. A typical benchmark task costs about $0.11 at default effort.
Is Mistral Small 3.2 24B free to use?›
There is no free API; the cheapest paid price is $0.075 per 1M input tokens. The weights are open, so you can run it yourself.
How can I use Mistral Small 3.2 24B?›
Through the API of DeepInfra, Featherless, Mistral, Parasail, Venice.
What is the context window of Mistral Small 3.2 24B?›
33K tokens, with up to 26K output tokens.
When was Mistral Small 3.2 24B released? Is there a newer version?›
Released 2025-06-20. It has been superseded by Mistral Small 4.
Is Mistral Small 3.2 24B open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run Mistral Small 3.2 24B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Mistral Small 3.2 24B read images?›
It accepts text and images as input.
What is Mistral Small 3.2 24B?›
Mistral Small 3.2 24B is an open-weights language model from Mistral, released Jun 2025.
What are the best alternatives to Mistral Small 3.2 24B?›
GPT-6.1 Terra (cheaper), GPT-6 Luna (better), Qwen3.8 27B (open weights).