Mistral
Devstral Small 2
active coding
Mistral's latest coding-focused model. Agentic coding capabilities.
Context 128K
Max output 8K
Pricing /1M tok $0.15 / $0.45 • API
Released Dec 1, 2025
// provider_directory
Mistral
All tracked Mistral models — pricing, context windows, lifecycle state, and release history in one place.
12 tracked models
9 live or preview
12 with listed pricing
Dec 1, 2025 latest release
// model_index
All Mistral models
Mistral
Magistral Medium 1.1
active reasoning
Mistral's flagship reasoning model with chain-of-thought capabilities. 128K context.
Context 128K
Max output 32K
Pricing /1M tok $2.00 / $6.00 • API
Released Jul 10, 2025
Mistral
Magistral Small 1.1
active reasoning
Smaller reasoning model. Open-source. 128K context.
Context 128K
Max output 32K
Pricing /1M tok $0.50 / $1.50 • API
Released Jul 10, 2025
Mistral
Devstral Medium 1.0
active coding
Medium-size coding model. Balanced capability and cost.
Context 128K
Max output 8K
Pricing /1M tok $0.80 / $2.40 • API
Released Jul 1, 2025
Mistral
Mistral Small 3.1
active open-weight
Open-weight small model. 24B parameters, 128K context, vision support.
Context 128K
Max output 8K
Pricing /1M tok $0.10 / $0.30 • OpenRouter
Released Mar 1, 2025
Mistral
Codestral 25.01
active coding
Mistral's coding specialist. 256K context, strong code generation.
Context 256K
Max output 8K
Pricing /1M tok $0.30 / $0.90 • API
Released Jan 1, 2025
Mistral
Mistral Large 2.1
active frontier
Mistral's largest non-reasoning model. 123B parameters, 128K context.
Context 128K
Max output 8K
Pricing /1M tok $2.00 / $6.00 • API
Released Nov 1, 2024
Mistral
Pixtral Large
active multimodal
Vision-capable variant of Mistral Large. 128K context, image understanding.
Context 128K
Max output 8K
Pricing /1M tok $2.00 / $6.00 • API
Released Nov 1, 2024
Lifecycle
← pixtral-12b
Mistral
Pixtral 12B
active open-weight
Open-weight 12B multimodal model. 128K context, image understanding.
Context 128K
Max output 8K
Pricing /1M tok $0.10 / $0.30 • OpenRouter
Released Sep 1, 2024
Lifecycle
→ pixtral-large
Mistral
Mistral Small 3.0
legacy open-weight
First Mistral Small 3 model. 24B params, text-only. Superseded by 3.1 with vision.
Context 128K
Max output 8K
Pricing /1M tok $0.10 / $0.30 • OpenRouter
Released Jan 1, 2025
Lifecycle
→ mistral-small-3-1
Mistral
Mixtral 8x7B
legacy open-weight
Pioneering open-weight MoE model. 8 experts, 7B each. 32K context.
Context 32K
Max output 8K
Pricing /1M tok $0.10 / $0.30 • OpenRouter
Released Dec 11, 2023
Mistral
Mistral 7B
legacy open-weight
Mistral's first open-weight model. 7B params, 32K context. Still used for local inference.
Context 32K
Max output 8K
Pricing /1M tok $0.02 / $0.06 • OpenRouter
Released Sep 27, 2023
// lineage_chains
Model lineage
Predecessor → successor chains tracked for Mistral models.
Pixtral 12B active Sep 1, 2024
→
Pixtral Large active Nov 1, 2024
Mistral Small 3.0 legacy Jan 1, 2025
→
Mistral Small 3.1 active Mar 1, 2025
// release_timeline
Mistral timeline
Devstral Small 2 released
Mistral's latest coding-focused model. Agentic coding capabilities.
Magistral Medium 1.1 released
Mistral's flagship reasoning model with chain-of-thought capabilities. 128K context.
Magistral Small 1.1 released
Smaller reasoning model. Open-source. 128K context.
Devstral Medium 1.0 released
Medium-size coding model. Balanced capability and cost.
Mistral Small 3.1 released
Open-weight small model. 24B parameters, 128K context, vision support.
Codestral 25.01 released
Mistral's coding specialist. 256K context, strong code generation.
Mistral Small 3.0 released
First Mistral Small 3 model. 24B params, text-only. Superseded by 3.1 with vision.
Mistral Large 2.1 released
Mistral's largest non-reasoning model. 123B parameters, 128K context.
Pixtral Large released
Vision-capable variant of Mistral Large. 128K context, image understanding.
Pixtral 12B released
Open-weight 12B multimodal model. 128K context, image understanding.
Mixtral 8x7B released
Pioneering open-weight MoE model. 8 experts, 7B each. 32K context.
Mistral 7B released
Mistral's first open-weight model. 7B params, 32K context. Still used for local inference.