Mistral: Models Intelligence, Performance & Price
Analysis of Mistral's models across key metrics including quality, price, output speed, latency, context window & more. This analysis is intended to support you in choosing the best model provided by Mistral for your use-case.
Most Intelligent
Intelligence index
Total 20 models
Fastest
Output speed
Total 20 models
Lowest Price
Blended price (per 1M tokens)
Total 20 models
Mistral offers 20 models, each with different intelligence, performance, and pricing characteristics. Below is a comparison of the key metrics across models.
- For intelligence, the top models on Mistral are Mistral Medium 3.5 (30), Mistral Small 4 (20), and Devstral 2 (19).
- For output speed, the fastest models are Ministral 3 3B (279 t/s), Mistral Small 4 (159 t/s), and Mistral Small 3.2 (150 t/s). Speed varies significantly across models, with a 93% difference between the fastest and slowest.
- For latency, Ministral 3 3B (0.64s), Mistral Small 3.2 (0.65s), and Ministral 3 8B (0.73s) offer the lowest time to first answer token.
- For pricing, Ministral 3 3B ($0.10), Mistral Small 3.1 ($0.12), and Mistral Small 3.2 ($0.12) offer the lowest blended prices per 1M tokens.
- For context window size, Mistral Medium 3.5 (262k), Devstral 2 (262k), and Mistral Large 3 (262k) support the largest context windows on Mistral.
- Ministral 3 3B offers both the fastest output and best pricing, making it attractive for throughput-sensitive and cost-conscious applications. Mistral Medium 3.5 leads in intelligence for tasks that require the highest quality.
Highlights
Intelligence Evaluations
Artificial Analysis Intelligence Index
Intelligence Evaluations
Agentic real-world work tasks, (Elo-500)/2000
Agentic tool use
Agentic coding & terminal use
Coding
Reasoning & knowledge
Scientific reasoning
Physics reasoning
Knowledge
1 - hallucination rate
Long context reasoning
Agentic knowledge work, Elo
Agentic SaaS workflows
Legal agentic work, criterion pass rate
Agentic business operations
Instruction following
Long-horizon agentic tasks
Kubernetes incident root-cause analysis
Visual reasoning
Intelligence Index vs. Price
Context Window
Context Window
Pricing
Intelligence Index vs. Price
Performance Summary
Output Speed vs. Price
Speed
Measured by Output Speed (tokens per second)
Output Speed
Latency
Measured by Time (seconds) to First Token
Latency: Time To First Answer Token
End-to-End Response Time
Seconds to output 500 tokens, calculated based on time to first token, 'thinking' time for reasoning models, and output speed
End-to-End Response Time vs. Price
Further Analysis | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
Mistral Medium 3.5 | 262k | Open | 30 | $0.61 | 66 | 2.16 | 40.15 | 30.39 | |||
Mistral Small 4 | 256k | Open | 20 | $0.10 | 159 | 0.79 | 16.51 | 12.57 | |||
Devstral 2 | 262k | Open | 19 | $0.00 | 50 | 1.38 | 11.37 | -- | |||
Magistral Medium 1.2 | 131k | Proprietary | 18 | $0.84 | 43 | 1.85 | 59.37 | 46.02 | |||
Devstral Small 2 | 256k | Open | 17 | $0.00 | 44 | 1.41 | 12.67 | -- | |||
Mistral Large 3 | 262k | Open | 16 | $0.08 | 49 | 1.10 | 11.20 | -- | |||
Mistral Small 3.1 | 131k | Open | 15 | $0.04 | 140 | 0.83 | 4.41 | -- | |||
Mistral Medium 3.1 | 131k | Proprietary | 15 | $0.16 | 86 | 1.51 | 7.33 | -- | |||
Mistral Medium 3 | 131k | Proprietary | 12* | -- | 48 | 1.51 | 11.96 | -- | |||
Mistral Small 4 | 262k | Open | 12* | -- | 141 | 0.84 | 4.38 | -- | |||
Magistral Small 1.2 | 131k | Open | 11 | $0.29 | 89 | 0.93 | 28.93 | 22.40 | |||
Ministral 3 14B | 262k | Open | 11 | $0.16 | 74 | 0.88 | 7.60 | -- | |||
Mistral Small 3.2 | 131k | Open | 11 | $0.14 | 150 | 0.65 | 3.98 | -- | |||
Ministral 3 8B | 262k | Open | 9 | $0.18 | 119 | 0.73 | 4.91 | -- | |||
Mistral Small 3 | 32.8k | Open | 7* | -- | 144 | 0.86 | 4.32 | -- | |||
Ministral 3 3B | 131k | Open | 7 | $0.13 | 279 | 0.64 | 2.43 | -- | |||
Mistral Small (Sep) | 32.8k | Open | 5* | -- | 146 | 0.83 | 4.25 | -- | |||
Mistral Small (Feb) | 262k | Proprietary | 4* | -- | 143 | 0.78 | 4.28 | -- | |||
Mistral Medium | 131k | Proprietary | 4* | -- | 46 | 2.17 | 12.92 | -- | |||
Mistral 7B | 32.8k | Open | 2* | -- | 112 | 0.81 | 5.29 | -- | |||
Key definitions
Frequently Asked Questions
Common questions about Mistral
Mistral offers 20 models that we track: Mistral Medium 3.5, Mistral Small 4, Devstral 2, Magistral Medium 1.2, Devstral Small 2, Mistral Large 3, Mistral Small 3.1, Mistral Medium 3.1, Mistral Medium 3, Mistral Small 4, Magistral Small 1.2, Ministral 3 14B, Mistral Small 3.2, Ministral 3 8B, Mistral Small 3, Ministral 3 3B, Mistral Small (Sep), Mistral Small (Feb), Mistral Medium, and Mistral 7B.
The most intelligent model available on Mistral is Mistral Medium 3.5 with an Intelligence Index score of 30.
The fastest model on Mistral by output speed is Ministral 3 3B at 279.4 tokens per second.
The model with the lowest time to first answer token on Mistral is Ministral 3 3B at 0.64s. Lower latency means faster initial response time.
The most affordable model on Mistral by blended price is Ministral 3 3B at $0.10 per 1M tokens (7:2:1 cache hit/input/output ratio).
Prices on Mistral vary up to 23x across models, from $0.10 per 1M tokens for Ministral 3 3B to $2.30 per 1M tokens for Magistral Medium 1.2.
Yes, Mistral offers an OpenAI-compatible API, making it easy to switch from OpenAI or use existing OpenAI SDK integrations.
19 of 20 models on Mistral support JSON mode for structured output.
Yes, all 20 models on Mistral support function calling (tool use).
Yes, Mistral offers 4 reasoning models: Mistral Medium 3.5, Mistral Small 4, Magistral Medium 1.2, and Magistral Small 1.2. Reasoning models use extended thinking to work through complex problems before providing an answer.
Yes, 15 of 20 models on Mistral are open weight models: Mistral Medium 3.5, Mistral Small 4, Devstral 2, Devstral Small 2, Mistral Large 3, Mistral Small 3.1, Mistral Small 4, Magistral Small 1.2, Ministral 3 14B, Mistral Small 3.2, Ministral 3 8B, Mistral Small 3, Ministral 3 3B, Mistral Small (Sep), and Mistral 7B.
Yes, provider performance can vary over time due to infrastructure changes, load balancing, and updates. We continuously benchmark all providers and display historical performance trends in the "Over Time" charts.
When choosing a model on Mistral, consider: intelligence (for quality-sensitive tasks), output speed (for throughput-intensive tasks), latency (for interactive applications requiring quick first responses), pricing (for cost-sensitive workloads), and features like context window size, JSON mode, or function calling support.