Enter opens the highlighted result. Esc closes.

Llama 3.1 8B Instruct vs MiniMax M2.1

MiniMax M2.1 leads Llama 3.1 8B Instruct on the Artificial Analysis Intelligence Index, 20.9 to 6.9, and Llama 3.1 8B Instruct costs 15.0x less per output token ($0.08 vs $1.20 per million). Updated October 22, 2026; 11 benchmarks have a result for both.

Replace Llama 3.1 8B Instruct Replace MiniMax M2.1
Llama 3.1 8B InstructMeta Llama
Coding
No data
Overall
17.9
Output $/M
$0.08
0
MiniMax M2.1MiniMax
Coding
No data
Overall
No data
Output $/M
$1.20
11
vs boards won of 11 shared

Llama 3.1 8B Instruct or MiniMax M2.1: which to pick

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

Pick Llama 3.1 8B Instruct for output tokens 15.0x cheaper ($0.08 vs $1.20 per million) and work measured by LiveCodeBench (Artificial Analysis run) or Terminal-Bench Hard (Artificial Analysis run), where it leads.

Model ALlama 3.1 8B InstructMeta LlamaNot benchmarkedNo coding benchmarkNo cost per solved taskModel BMiniMax M2.1MiniMaxNot benchmarked1 independent benchmark$0.59 per solved task
Shared results
11
Llama 3.1 8B Instruct only
40
MiniMax M2.1 only
23
Like-for-like categories
7 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
Llama 3.1 8B Instruct not benchmarked, MiniMax M2.1 not benchmarked.
Widest category
Math, 78.4 points apart on average over 2 shared benchmarks.
Widest single gap
AA Math Index: Llama 3.1 8B Instruct 4.3 against MiniMax M2.1 82.7, 78.4 points apart on a 0 to 100 scale.
Output price
Llama 3.1 8B Instruct $0.08 against MiniMax M2.1 $1.20 per million output tokens, 15.0x apart.
Context window
Llama 3.1 8B Instruct takes 131K tokens, MiniMax M2.1 205K.

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
CategoryLlama 3.1 8B InstructBarsMiniMax M2.1Ties
Coding02
Intelligence01
Agentic work01
Reasoning02
Math02
Knowledge01
General02
All shared boards011
CodingLike-for-likeSix of these boards feed the Vibe Coding Index02
Coding benchmarks for Llama 3.1 8B Instruct and MiniMax M2.1
BenchmarkLlama 3.1 8B InstructMiniMax M2.1Leader
AA Coding Index5.4Not measuredOnly Llama 3.1 8B Instruct measured
LiveCodeBench (Vals AI)Not measured81.8%Only MiniMax M2.1 measured
SWE-bench Verified (Vals AI)Not measured74.8%Only MiniMax M2.1 measured
LiveCodeBench (Artificial Analysis run)11.6%81.0%MiniMax M2.1
Terminal-Bench Hard (Artificial Analysis run)0.8%28.8%MiniMax M2.1
Terminal-BenchNot measured36.6%Only MiniMax M2.1 measured
WeirdML1.7%Not measuredOnly Llama 3.1 8B Instruct measured
ALE-BenchNot measured623.8Only MiniMax M2.1 measured
LMArena Text: coding1195Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: software and IT1195Not measuredOnly Llama 3.1 8B Instruct measured
Neither model has a result on the other 33 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
Intelligence benchmarks for Llama 3.1 8B Instruct and MiniMax M2.1
BenchmarkLlama 3.1 8B InstructMiniMax M2.1Leader
AA Intelligence Index6.920.9MiniMax M2.1
Epoch Capabilities Index116.5Not measuredOnly Llama 3.1 8B Instruct measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for Llama 3.1 8B Instruct and MiniMax M2.1
BenchmarkLlama 3.1 8B InstructMiniMax M2.1Leader
τ²-Bench Telecom16.4%85.4%MiniMax M2.1
BALROG15.1%Not measuredOnly Llama 3.1 8B Instruct measured
Berkeley Function Calling Leaderboard (BFCL)25.8%Not measuredOnly Llama 3.1 8B Instruct measured
Neither model has a result on the other 39 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index02
Reasoning benchmarks for Llama 3.1 8B Instruct and MiniMax M2.1
BenchmarkLlama 3.1 8B InstructMiniMax M2.1Leader
GPQA DiamondNot measured78.5%Only MiniMax M2.1 measured
GPQA Diamond (Artificial Analysis run)25.9%83.0%MiniMax M2.1
Humanity's Last Exam (Artificial Analysis run)5.3%23.2%MiniMax M2.1
GPQA Diamond (Epoch AI run)27.0%Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: hard prompts1175Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: expert1144Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: hard prompts in English1196Not measuredOnly Llama 3.1 8B Instruct measured
Neither model has a result on the other 16 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index02
Math benchmarks for Llama 3.1 8B Instruct and MiniMax M2.1
BenchmarkLlama 3.1 8B InstructMiniMax M2.1Leader
AIMENot measured77.9%Only MiniMax M2.1 measured
MATH 500 (Vals AI)Not measured89.0%Only MiniMax M2.1 measured
MGSM (Vals AI)Not measured87.8%Only MiniMax M2.1 measured
AA Math Index4.382.7MiniMax M2.1
AIME 2025 (Artificial Analysis run)4.3%82.7%MiniMax M2.1
OTIS Mock AIME1.7%Not measuredOnly Llama 3.1 8B Instruct measured
MATH level 522.9%Not measuredOnly Llama 3.1 8B Instruct measured
GSM8K82.4%Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: math1179Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: mathematical work1180Not measuredOnly Llama 3.1 8B Instruct measured
Neither model has a result on the other 4 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index01
Knowledge benchmarks for Llama 3.1 8B Instruct and MiniMax M2.1
BenchmarkLlama 3.1 8B InstructMiniMax M2.1Leader
MMLU ProNot measured87.0%Only MiniMax M2.1 measured
LegalBench (Vals AI)Not measured75.5%Only MiniMax M2.1 measured
CaseLaw (Vals AI)Not measured55.8%Only MiniMax M2.1 measured
CorpFin (Vals AI)Not measured58.9%Only MiniMax M2.1 measured
TaxEval (Vals AI)Not measured66.3%Only MiniMax M2.1 measured
MedQA (Vals AI)Not measured91.2%Only MiniMax M2.1 measured
MedCode (Vals AI)Not measured34.1%Only MiniMax M2.1 measured
MMLU Pro (Artificial Analysis run)47.6%87.5%MiniMax M2.1
MMLU56.1%Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: science1193Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: medicine and healthcare1182Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: legal and government1206Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: business and finance1166Not measuredOnly Llama 3.1 8B Instruct measured
Neither model has a result on the other 7 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
Web design benchmarks for Llama 3.1 8B Instruct and MiniMax M2.1
BenchmarkLlama 3.1 8B InstructMiniMax M2.1Leader
Design Arena: websitesNot measured1201Only MiniMax M2.1 measured
Design Arena: UI componentsNot measured1220Only MiniMax M2.1 measured
Design Arena: game devNot measured1134Only MiniMax M2.1 measured
Design Arena: data vizNot measured1206Only MiniMax M2.1 measured
Design Arena: 3DNot measured1164Only MiniMax M2.1 measured
Design Arena: all code categoriesNot measured1190Only MiniMax M2.1 measured
Design Arena: SVGNot measured1131Only MiniMax M2.1 measured
Neither model has a result on the other 2 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index02
General benchmarks for Llama 3.1 8B Instruct and MiniMax M2.1
BenchmarkLlama 3.1 8B InstructMiniMax M2.1Leader
MedScribe (Vals AI)Not measured80.8%Only MiniMax M2.1 measured
IFBench28.6%69.9%MiniMax M2.1
Long context reasoning (AA-LCR)18.0%67.7%MiniMax M2.1
DTBench50.9%Not measuredOnly Llama 3.1 8B Instruct measured
LMCA5.4%Not measuredOnly Llama 3.1 8B Instruct measured
PIQA81.2%Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text1187Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: instruction following1159Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: creative writing1154Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: English1218Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: non-English1148Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: Chinese1151Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: German1144Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: Japanese1061Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: Korean1053Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: French1178Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: Spanish1169Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: Russian1158Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: multi-turn1172Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: longer queries1181Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: writing and language1162Not measuredOnly Llama 3.1 8B Instruct measured
LMArena Text: entertainment and media1145Not measuredOnly Llama 3.1 8B Instruct measured
Neither model has a result on the other 26 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Llama 3.1 8B Instruct and MiniMax M2.1
BenchmarkLlama 3.1 8B InstructMiniMax M2.1Leader
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

Llama 3.1 8B Instruct$0.0001Fits in one request
MiniMax M2.1$0.0009Fits in one request

Llama 3.1 8B Instruct costs less

Repository review

50K input + 3K output tokens

Llama 3.1 8B Instruct$0.0027Fits in one request
MiniMax M2.1$0.019Fits in one request

Llama 3.1 8B Instruct costs less

Agent loop

200K input + 10K output tokens

Llama 3.1 8B Instruct$0.011Exceeds the context window
MiniMax M2.1$0.072Exceeds the context window

Llama 3.1 8B Instruct 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 Llama 3.1 8B Instruct and MiniMax M2.1
SpecificationMeta Llama 3.1 8B InstructMiniMax M2.1
MakerMeta LlamaMiniMax
WeightsOpen weightsOpen weights
ReleasedJuly 23, 2024December 23, 2025
Context window131K205K
Input $/Mtok$0.05$0.30
Output $/Mtok$0.08$1.20
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it01
Cost per solved taskNot computable$0.59
SourceOpenRouter listingOpenRouter listing

Questions

Is Llama 3.1 8B Instruct better than MiniMax M2.1 for coding?

Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Llama 3.1 8B Instruct for output tokens 15.0x cheaper ($0.08 vs $1.20 per million) and work measured by LiveCodeBench (Artificial Analysis run) or Terminal-Bench Hard (Artificial Analysis run), where it leads.

Which is cheaper, Llama 3.1 8B Instruct or MiniMax M2.1?

Llama 3.1 8B Instruct costs $0.05 per million input tokens and $0.08 per million output tokens; MiniMax M2.1 costs $0.30 and $1.20. Cost per solved task is known for MiniMax M2.1 only.

Which has the bigger context window, Llama 3.1 8B Instruct or MiniMax M2.1?

MiniMax M2.1, with 205K tokens against 131K.

How many benchmarks is the Llama 3.1 8B Instruct vs MiniMax M2.1 comparison based on?

11 benchmarks have a result for both models as of October 22, 2026. Llama 3.1 8B Instruct's coding index rests on 0 independent benchmarks and MiniMax M2.1's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next