LLM Pricing
PricingLeaderboardsToolsProvidersReleasesGuides

© 2026 LLM Pricing

About
·Contact
·Privacy
·RSS
← Model list

Devstral 2

mistral·mistral/devstral-2512·Deprecated·Open weights·devstral series
Mistral's coding-agent model for repository work, terminal tasks, and software fixes

Devstral 2 by mistral is offered by 13 providers on this page. Public prices are shown for 12 of them. Its official list price is $0.40 per 1M input tokens and $2.00 per 1M output tokens. The lowest paid channel is Kilo Gateway at $0.40 / $2.00 per 1M, about 1.4× below the list price. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it.

Artificial Analysis rates it 8.6 on the Intelligence Index, with 31.3 for coding and 2.3 for agentic tasks. Against its lowest blended price of $0.80 per 1M, that is roughly 11 index points per dollar, which is the value ratio the leaderboards rank on.

The context window is 262,144 tokens at the reference host, but hosts report different limits, from 200,000 to 262,144, so the usable window depends on the provider you pick. It supports tool use and structured output. Accepted input modalities are Text and PDF. The weights are open, so it can also be self-hosted or served through a gateway of your choice. Providers report the capability flags inconsistently, so verify a specific feature against the host you plan to use. Its training knowledge cuts off at 2025-12.

Specs & pricing

Input / output per 1M tokens
Official price·Mistral
$0.40 / $2.00
Blended $0.80 · Cache read —
Lowest paid·Kilo GatewayGateway
$0.40 / $2.00
Blended $0.80 · 1.4× spread
Context
262,144
Output limit
262,144
Knowledge cutoff
2025-12
Released / updated
2025-12-09 / 2025-12-09
Capabilities
Reasoning✓ Tool use✓ Structured output✓ TemperatureAttachments
⚠ Providers report capability flags inconsistently
Modalities
TextPDF
Weights
Hugging Face

Quality & performanceArtificial Analysis · Intelligence Index v4.3

Intelligence8.6
Value11
Coding31.3
Agentic2.3
Output speed—
TTFT—
Value formulaIQ 8.6 ÷ min blended $0.800 = 11

Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.

Available at 13 providers12 with public prices

ProviderTierInputOutputCache readCache writeContextOutput limitStatus
Kilo GatewayGateway$0.40$2.00$0.04—262,144209,715
OpenRouterGateway$0.40$2.00$0.04—262,144209,715
Merge GatewayGateway$0.40$2.00$0.04—256,000 ⚠256,000
OpperGateway$0.40$2.00——256,000 ⚠8,192
Pioneer
devstral-2
Gateway$0.40$2.00$0.40$0.40256,000 ⚠131,072
Eden AIGateway$0.40$2.00$0.04—262,144262,144
MistralOfficialFirst-party$0.40$2.00——262,144262,144deprecated
MistralOfficial
devstral-latest
First-party$0.40$2.00——262,144262,144deprecated
DevPass (LLM Gateway)Gateway$0.40$2.00——262,144262,144deprecated
LLM GatewayGateway$0.40$2.00——262,144262,144

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Your usage cost

1Kilo Gateway$136.80
2OpenRouter$136.80
3Merge Gateway$136.80
4Eden AI$136.80
5Cortecs$163.24
6Opper$180.00
8Mistral · Official$180.00
Switch to Kilo Gateway to save $43.20/mo (24%)
Note: this is a gateway; verify availability and rate limits yourself.

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$2.00$02026-08-052026-08-132026-08-05 · Input list $0.402026-08-06 · Input list $0.402026-08-07 · Input list $0.402026-08-08 · Input list $0.402026-08-09 · Input list $0.402026-08-10 · Input list $0.402026-08-11 · Input list $0.402026-08-12 · Input list $0.402026-08-13 · Input list $0.402026-08-05 · Output list $2.002026-08-06 · Output list $2.002026-08-07 · Output list $2.002026-08-08 · Output list $2.002026-08-09 · Output list $2.002026-08-10 · Output list $2.002026-08-11 · Output list $2.002026-08-12 · Output list $2.002026-08-13 · Output list $2.002026-08-05 · Min blended $0.802026-08-06 · Min blended $0.802026-08-07 · Min blended $0.802026-08-08 · Min blended $0.802026-08-09 · Min blended $0.802026-08-10 · Min blended $0.802026-08-11 · Min blended $0.802026-08-12 · Min blended $0.802026-08-13 · Min blended $0.80

Benchmark3 items

NameConditionsScoreMetricSource
Artificial Analysis Coding Index—23.7indexSource ↗
SciCode—33.1percent correctSource ↗
Terminal-Bench Hard—18.9success rateSource ↗

The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.

Artificial Analysis evaluations12 items

GPQA Diamond59.4%
Humanity's Last Exam3.6%
MMLU-Pro76.2%
SciCode32.8%
LiveCodeBench44.8%
Terminal-Bench Hard18.9%
Terminal-Bench 2.130.3%
AIME 202536.7%
τ²-Bench Telecom24.9%
τ³-Bench Banking10.5%
AA-LCR32.3%
IFBench38.1%

Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and harness, so rows are not comparable with one another. The Intelligence Index above draws on a different, newer set of evaluations.

Related models

devstral-latestsame series$0.44 / $2.20devstral-latest@eusame series$0.44 / $2.20Devstral 2 123Bsame series$0.40 / $2.00Gemma 3 12B ITcheaper alternative$0.27 / $0.27Qwen3 30B A3B Instruct 2507cheaper alternative$0.30 / $0.50DeepSeek V3.2 Expcheaper alternative$0.29 / $0.43Grok 4.1 Fast (Non-Reasoning)cheaper alternative$0.20 / $0.50
Data partly from models.dev (MIT) · Mistral official docs ↗