Enter opens the highlighted result. Esc closes.

Devstral 2 2512 vs Granite 4.2 8B

Granite 4.2 8B leads Devstral 2 2512 on the Artificial Analysis Intelligence Index, 11.1 to 8.6, and Granite 4.2 8B costs 10.0x less per output token ($0.20 vs $2.00 per million). Updated October 22, 2026; 3 benchmarks have a result for both.

Replace Devstral 2 2512 Replace Granite 4.2 8B
Devstral 2 2512Mistral AI
Coding
No data
Overall
No data
Output $/M
$2.00
2
Granite 4.2 8BIBM Granite
Coding
No data
Overall
No data
Output $/M
$0.20
1
vs boards won of 3 shared

Devstral 2 2512 or Granite 4.2 8B: which to pick

Neither model has a coding score yet, so the two cannot be ranked on coding.

Pick Devstral 2 2512 for a larger context window (262K against 131K tokens) and work measured by AA Intelligence Index, where it leads.

Model ADevstral 2 2512Mistral AINot benchmarked1 independent benchmark$1.19 per solved taskModel BGranite 4.2 8BIBM GraniteNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
3
Devstral 2 2512 only
3
Granite 4.2 8B only
24
Like-for-like categories
3 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
Devstral 2 2512 not benchmarked, Granite 4.2 8B not benchmarked.
Widest category
Coding, 8.9 points apart on average over 1 shared benchmark.
Widest single gap
AA Coding Index: Devstral 2 2512 31.3 against Granite 4.2 8B 22.4, 8.9 points apart on a 0 to 100 scale.
Output price
Devstral 2 2512 $2.00 against Granite 4.2 8B $0.20 per million output tokens, 10.0x apart.
Context window
Devstral 2 2512 takes 262K tokens, Granite 4.2 8B 131K.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryDevstral 2 2512BarsGranite 4.2 8BTies
Coding10
Intelligence01
Agentic work10
All shared boards21
CodingLike-for-likeSix of these boards feed the Vibe Coding Index10
Coding benchmarks for Devstral 2 2512 and Granite 4.2 8B
BenchmarkDevstral 2 2512Granite 4.2 8BLeader
SWE-bench Verified53.8%Not measuredOnly Devstral 2 2512 measured
AA Coding Index31.322.4Devstral 2 2512
LiveCodeBench (Vals AI)51.8%Not measuredOnly Devstral 2 2512 measured
SWE-bench Verified (Vals AI)62.8%Not measuredOnly Devstral 2 2512 measured
SciCodeNot measured31.5%Only Granite 4.2 8B measured
LMArena Text: codingNot measured1378Only Granite 4.2 8B measured
LMArena Text: software and ITNot measured1367Only Granite 4.2 8B measured
Neither model has a result on the other 32 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
Intelligence benchmarks for Devstral 2 2512 and Granite 4.2 8B
BenchmarkDevstral 2 2512Granite 4.2 8BLeader
AA Intelligence Index8.611.1Granite 4.2 8B
Neither model has a result on the other 1 board here
Agentic workLike-for-likePublic boards, shown as published, kept out of the index10
Agentic work benchmarks for Devstral 2 2512 and Granite 4.2 8B
BenchmarkDevstral 2 2512Granite 4.2 8BLeader
AA Agentic Index2.31.3Devstral 2 2512
Neither model has a result on the other 40 boards here
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index5 boards, none shared
Reasoning benchmarks for Devstral 2 2512 and Granite 4.2 8B
BenchmarkDevstral 2 2512Granite 4.2 8BLeader
GPQA Diamond (Artificial Analysis run)Not measured63.1%Only Granite 4.2 8B measured
Humanity's Last Exam (Artificial Analysis run)Not measured9.7%Only Granite 4.2 8B measured
LMArena Text: hard promptsNot measured1327Only Granite 4.2 8B measured
LMArena Text: expertNot measured1386Only Granite 4.2 8B measured
LMArena Text: hard prompts in EnglishNot measured1345Only Granite 4.2 8B measured
Neither model has a result on the other 17 boards here
MathNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Math benchmarks for Devstral 2 2512 and Granite 4.2 8B
BenchmarkDevstral 2 2512Granite 4.2 8BLeader
Neither model has a result on 13 boards here
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index4 boards, none shared
Knowledge benchmarks for Devstral 2 2512 and Granite 4.2 8B
BenchmarkDevstral 2 2512Granite 4.2 8BLeader
LMArena Text: scienceNot measured1336Only Granite 4.2 8B measured
LMArena Text: medicine and healthcareNot measured1336Only Granite 4.2 8B measured
LMArena Text: legal and governmentNot measured1327Only Granite 4.2 8B measured
LMArena Text: business and financeNot measured1322Only Granite 4.2 8B measured
Neither model has a result on the other 14 boards here
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Web design benchmarks for Devstral 2 2512 and Granite 4.2 8B
BenchmarkDevstral 2 2512Granite 4.2 8BLeader
Neither model has a result on 9 boards here
GeneralDirectional onlyPublic boards, shown as published, kept out of the index12 boards, none shared
General benchmarks for Devstral 2 2512 and Granite 4.2 8B
BenchmarkDevstral 2 2512Granite 4.2 8BLeader
Long context reasoning (AA-LCR)Not measured45.0%Only Granite 4.2 8B measured
LMArena TextNot measured1319Only Granite 4.2 8B measured
LMArena Text: instruction followingNot measured1300Only Granite 4.2 8B measured
LMArena Text: creative writingNot measured1236Only Granite 4.2 8B measured
LMArena Text: EnglishNot measured1337Only Granite 4.2 8B measured
LMArena Text: non-EnglishNot measured1301Only Granite 4.2 8B measured
LMArena Text: ChineseNot measured1367Only Granite 4.2 8B measured
LMArena Text: RussianNot measured1282Only Granite 4.2 8B measured
LMArena Text: multi-turnNot measured1301Only Granite 4.2 8B measured
LMArena Text: longer queriesNot measured1325Only Granite 4.2 8B measured
LMArena Text: writing and languageNot measured1269Only Granite 4.2 8B measured
LMArena Text: entertainment and mediaNot measured1271Only Granite 4.2 8B measured
Neither model has a result on the other 34 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Devstral 2 2512 and Granite 4.2 8B
BenchmarkDevstral 2 2512Granite 4.2 8BLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

Devstral 2 2512$0.0014Fits in one request
Granite 4.2 8B$0.0002Fits in one request

Granite 4.2 8B costs less

Repository review

50K input + 3K output tokens

Devstral 2 2512$0.026Fits in one request
Granite 4.2 8B$0.0046Fits in one request

Granite 4.2 8B costs less

Agent loop

200K input + 10K output tokens

Devstral 2 2512$0.100Fits in one request
Granite 4.2 8B$0.018Exceeds the context window

Granite 4.2 8B costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of Devstral 2 2512 and Granite 4.2 8B
SpecificationMistral Devstral 2 2512IBM Granite 4.2 8B
MakerMistral AIIBM Granite
WeightsOpen weightsOpen weights
ReleasedDecember 9, 2025August 31, 2026
Context window262K131K
Input $/Mtok$0.40$0.08
Output $/Mtok$2.00$0.20
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it10
Cost per solved task$1.19Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Devstral 2 2512 better than Granite 4.2 8B for coding?

Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Devstral 2 2512 for a larger context window (262K against 131K tokens) and work measured by AA Intelligence Index, where it leads.

Which is cheaper, Devstral 2 2512 or Granite 4.2 8B?

Devstral 2 2512 costs $0.40 per million input tokens and $2.00 per million output tokens; Granite 4.2 8B costs $0.08 and $0.20. Cost per solved task is known for Devstral 2 2512 only.

Which has the bigger context window, Devstral 2 2512 or Granite 4.2 8B?

Devstral 2 2512, with 262K tokens against 131K.

How many benchmarks is the Devstral 2 2512 vs Granite 4.2 8B comparison based on?

3 benchmarks have a result for both models as of October 22, 2026. Devstral 2 2512's coding index rests on 1 independent benchmark and Granite 4.2 8B's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next