Devstral Medium 2507

MistralReleasedJul 10, 2025KnowledgeMay 2025
Context
128K
Max output
128K
Serving providers
1
Cheapest input
$0.40 /1M
Prices are on-demand list rates. Enterprise commitments (committed-use discounts, provisioned throughput, negotiated rates) can be materially lower.

Benchmarks

independent evaluations · source: Artificial Analysis
Speed & latency
Output speed
75 t/s
tokens / second
Time to first token
8.90s
latency
Headline indices
Individual benchmarks · click a row for effort variants
BenchmarkBest scoreEffortCost / task
CritPt0%default
GPQA Diamond49%default
Humanity's Last Exam4%default
IFBench30%default
LiveCodeBench34%default
Long-Context Reasoning29%default
MMLU-Pro71%default
Omniscience-31.6default
SciCode29%default
TerminalBench Hard9%default
τ²-bench20%default

Pricing by serving provider

standard tier · per 1M tokens · click a provider to expand
Serving providerInput /1MOutput /1MEndpoints
MistralCheapest$0.40$2.001
ComponentUnitStandardBatchCached
Text input/1M tok$0.4000
Text output/1M tok$2.00

Price history

input + output $/1M since we started tracking
Input Output
$0.000$0.500$1.00$1.50$2.00Jul 10Jun 22$2.00$0.400

Cost calculator

estimate your monthly spend on this model
$240
estimated / month

Model IDs

copy the exact identifier for your platform
mistral/devstral-medium-2507

Frequently asked questions

Devstral Medium 2507 pricing, context and availability

How much does Devstral Medium 2507 cost?

Devstral Medium 2507 costs $0.40 per 1M input tokens and $2.00 per 1M output tokens at its cheapest provider via Mistral.

What is the context window of Devstral Medium 2507?

Devstral Medium 2507 accepts up to 128K tokens of context and can return up to 128K output tokens.

What can Devstral Medium 2507 do?

Devstral Medium 2507 supports tool use and structured output.

Other Mistral models

compare pricing across the Mistral lineup
Pricing verified against LiteLLM + OpenRouter + provider pages · refreshed daily

Every weekday

AI moves fast. Here's your debrief.

News, analysis, tools, and more.

For people who build with AI