VerdictPal · editorial desk · 2026VerdictPal

Compare · Two models, head-to-head

Pick two models. See the diff.

Composite scores, benchmark ranks, capability grids, and pricing per million tokens. All sourced from the same data that powers the model atlas.

Cards Models

Mistral Large 3 vs Llama 4 Maverick

Model A

Model B

MODEL A

Mistral Large 3

Mistral · Mistral Large

Composite 48Intel 1620 rows6 sources

Mistral's December 2025 flagship MoE. 262K context, vision input, $0.50/$1.50 per 1M tokens. 675B total / 41B active parameters.

MODEL B

Llama 4 Maverick

Meta · Llama 4

Composite 49Intel 1421 rows6 sources

Meta's reasoning-focused Llama 4 MoE. 1M context, 17B active / 400B total params, 128 experts. Open weights for multimodal assistants and coding.

Mistral Large 32stats won

Tied0stats

Llama 4 Maverick5stats won

Spec showdown

Capacity, price, freshness.

Mistral Large 3vsLlama 4 Maverick

Composite scores are within 5 points — 48 for Mistral Large 3 vs 49 for Llama 4 Maverick. The choice comes down to capability profile and price. Mistral Large 3 has the higher Vals/AA intelligence score (16 vs 14). Llama 4 Maverick is more affordable at $0.17/1M input — about 3× cheaper than $0.50/1M. Llama 4 Maverick's 1M context window dwarfs Mistral Large 3's 262K — a meaningful difference for long-document or codebase work.

Composite score
Mistral Large 348Llama 4 Maverick49Winner on this row
Intelligence (Vals/AA)
Mistral Large 316Winner on this rowLlama 4 Maverick14
Context window
Mistral Large 3262KLlama 4 Maverick1MWinner on this row
Max output
Mistral Large 38KLlama 4 Maverick16KWinner on this row
Input price / 1M (lower wins)
Mistral Large 3$0.50Llama 4 Maverick$0.17Winner on this row
Output speed (AA default API)
Mistral Large 364 t/sLlama 4 Maverick110 t/sWinner on this row
Time to first token (AA default API)
Mistral Large 30.59sWinner on this rowLlama 4 Maverick0.60s

Composite score

Intelligence (Vals/AA)

Context window

262K

Max output

16K

Input price / 1M (lower wins)

$0.50

$0.17

Output speed (AA default API)

64 t/s

110 t/s

Time to first token (AA default API)

0.59s

0.60s

Capability grid

What each model can do.

Modalities

Mistral Large 3Llama 4 Maverick

Text●●

Image●●

Audio——

Video——

Code●●

Tool use●●

Reasoning●●

Access

Mistral Large 3Llama 4 Maverick

API●●

Consumer app——

Open weights●●

Managed cloud●●

On-device——

Family radar

Top score per benchmark family.

Knowledge

Reasoning

Math

Coding

Agentic

Long context

Tool use

Benchmark ranks

Every shared benchmark, side by side.

Knowledge

MMLU#13/22 · #19/22

87.8%

86.1%

MMLU-Pro#18/21 · #11/21

72.8%

80.9%

Reasoning

Artificial Analysis Intelligence Index#58/67 · #59/67

15.9%

14.3%

GPQA Diamond#62/69 · #63/69

71.2%

69.4%

Humanity's Last Exam#64/65 · #61/65

4.1%

4.8%

IFBench#57/57 · #53/57

36.19%

42.99%

LiveBench#11/21 · #13/21

62.8%

62.1%

Math

AIME#15/19 · #16/19

38%

19.33%

Coding

HumanEval#12/15 · #14/15

92.4% pass@1

91.8% pass@1

LiveCodeBench#9/13 · #10/13

46.5%

39.7%

SciCode#55/65 · #58/65

36.2%

33.1%

SWE-bench Verified#21/23 · #22/23

76.2%

74.8%

Agentic

Terminal-Bench#64/69 · #66/69

11.99%

7.87%

τ³-Bench Banking#61/64 · #63/64

5.77%

3.92%

Long context

AA-LCR#57/64 · #56/64

34.67%

46%

LongBench v2#11/19 · #12/19

50.8%

49.8%

Tool use

BFCL#8/20 · #15/20

80.1%

75.9%

Performance

AA output speed#44/69 · #24/69

64 t/s

110 t/s

AA time to first token#55/69 · #54/69

0.59s

0.60s

Dossier facts

Dates, sources, fineprint.

Mistral Large 3Llama 4 Maverick

ProviderMistralMeta

FamilyMistral LargeLlama 4

Release date2025-12-022026-04-05

Knowledge cutoff2025-06-012024-08-01

Context window262K1M

Max output8K16K

Input price /1M$0.50$0.17

Output price /1M$1.5$0.60

Statussolidflagship

Composite score4849

Intelligence score1614

Intel sourceaaaa

Benchmark rows2021

Source count66

Freshnessfreshfresh

Pricing checked2026-06-092026-06-09

Ecosystem

Which tools wrap each model.

Mistral Large 3

mistral le chat

Llama 4 Maverick

Not listed in any tool yet.

Verdict

Which model to actually pick.

Open Mistral Large 3's dossier Open Llama 4 Maverick's dossier