Devstral Small (Jul '25)
Jul 2025
Devstral Small (Jul '25)Compare inon
Benchmarks of Devstral Small (Jul '25)7
BenchmarkByScoreRank
Legacy · saturated or no longer reported1
No benchmark matches
Rankingconfiguration unstated?
Retired models are left out of the current rankings.
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperGPT OSS 120BhighOpen×3.2-60%$0.11Context: 256K → 131K+1BetterGPT OSS 20BhighOpen×3-76%$0.011Context: 256K → 131K
OpenNemotron 3 Nano 30B A3Bthinking×3.1-30%$0.017Context: 256K → 131K+1CheaperGPT-5 Nanohigh×2.9-8%$0.05 / $0.40CheaperNVIDIA Nemotron Nano 9B V2thinkingOpen×2.4-53%$0.04 / $0.16Context: 256K → 131KModelScoreCostPer taskWhat changes
GPT OSS 120BhighOpen×3.2-60%$0.11Context: 256K → 131K+1Nemotron 3 Nano 30B A3BthinkingOpen×3.1-30%$0.017Context: 256K → 131K+1GPT OSS 20BhighOpen×3-76%$0.011Context: 256K → 131KGPT-5 Nanohigh×2.9-8%$0.05 / $0.40NVIDIA Nemotron Nano 9B V2thinkingOpen×2.4-53%$0.04 / $0.16Context: 256K → 131KModelScoreCostPer taskWhat changes
GPT OSS 20BhighOpen×3-76%$0.011Context: 256K → 131KMinistral 3 8B 2512Open+8%same$0.011Nemotron 3 Nano 30B A3BthinkingOpen×3.1-30%$0.017Context: 256K → 131K+1
DeepSeek V3 0324non-thinkingOpen+40%+170%$0.021Context: 256K → 128K+1Mistral Large 3non-thinkingOpen+30%+400%$0.031Memory (4-bit): not verifiedModelScoreCostPer taskWhat changes
Nemotron 3 Nano 30B A3BthinkingOpen×3.1-30%$0.017Context: 256K → 131K+1GPT OSS 20BhighOpen×3-76%$0.011Context: 256K → 131KSApriel v1.6 15B ThinkerOpen×3——Context: 256K → 128KSApriel v1.5 15B ThinkerOpen×3——Context: 256K → 128K
Qwen3 VL 32B (Reasoning)Open×2.9+250%$0.20 / $1.50Memory (4-bit): 14 GB → 20 GBModelScoreCostPer taskWhat changes
Magistral Small 1thinkingOpen+41%—$0.70 / $1.16Context: 256K → 128KMagistral Medium 1thinking+38%——Context: 256K → 40KMistral Large 3non-thinkingOpen+30%—$0.031Memory (4-bit): not verifiedMinistral 3 8B 2512Open+8%—$0.011Ministral 3 14B 2512non-thinkingOpensame—$0.019Ministral 3 3B 2512Open-25%—$0.008Context: 256K → 131KDevstral Mediumnon-thinkingOpen-84%—$0.40 / $2Context: 256K → 128K+1Mistral Small 3non-thinkingOpen-85%—$0.05 / $0.08Context: 256K → 32KAll Mistral models →Run it yourself
estimateFits on✗ 8 GB (phone, laptop)✓ 16 GB · compressed✓ 24 GB GPU · compressed✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceApache-2.0free for commercial use
History
3 versionsBenchmarks
all categoriesCoding
SWE-rebench · all · January 1, 2025 to August 1, 202514.6
SWE-rebench · Go · January 1, 2025 to August 1, 20250
SWE-rebench · Java · January 1, 2025 to August 1, 20250
SWE-rebench · Rust · January 1, 2025 to August 1, 20250
SWE-rebench · Python · January 1, 2025 to August 1, 202514.6
SWE-rebench · TypeScript · January 1, 2025 to August 1, 20250
FAQ
How much does Devstral Small (Jul '25) cost?›
$0.10 per 1M input tokens and $0.30 per 1M output tokens.
Is Devstral Small (Jul '25) free to use?›
There is no free API; the cheapest paid price is $0.10 per 1M input tokens. The weights are open, so you can run it yourself.
How can I use Devstral Small (Jul '25)?›
Through the API of Mistral.
What is the context window of Devstral Small (Jul '25)?›
256K tokens, with up to 128K output tokens.
When was Devstral Small (Jul '25) released? Is there a newer version?›
Released 2025-07-10. It has been superseded by Devstral Small 2.
Is Devstral Small (Jul '25) open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run Devstral Small (Jul '25) locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Devstral Small (Jul '25) read images?›
No, it accepts text only.
What is Devstral Small (Jul '25)?›
Devstral Small (Jul '25) is an open-weights language model from Mistral, released Jul 2025, the successor to Devstral Small (May '25).
What are the best alternatives to Devstral Small (Jul '25)?›
GPT OSS 120B (cheaper), GPT OSS 20B (better), Nemotron 3 Nano 30B A3B (open weights).