Compare · ModelsLive · 2 picked · head to head
Mistral Large vs Grok-2 (Dec 2024)
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Grok-2 (Dec 2024) wins on 5/7 benchmarks
Grok-2 (Dec 2024) wins 5 of 7 shared benchmarks. Leads in math · knowledge · reasoning.
Category leads
coding·Mistral Largemath·Grok-2 (Dec 2024)knowledge·Grok-2 (Dec 2024)reasoning·Grok-2 (Dec 2024)
Hype vs Reality
Attention vs performance
Mistral Large
#186 by perf·no signal
Grok-2 (Dec 2024)
#176 by perf·no signal
Vendor risk
Who is behind the model
Mistral AI
$14.0B·Tier 1
xAI
$250.0B·Tier 1
Head to head
7 benchmarks · 2 models
Mistral LargeGrok-2 (Dec 2024)
Aider · Code Editing
Mistral Large leads by +1.6
Mistral Large
60.2
Grok-2 (Dec 2024)
58.6
FrontierMath-2025-02-28-Private
Grok-2 (Dec 2024) leads by +0.3
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
Mistral Large
0.3
Grok-2 (Dec 2024)
0.7
GPQA diamond
Grok-2 (Dec 2024) leads by +20.0
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Mistral Large
18.4
Grok-2 (Dec 2024)
38.4
Lech Mazur Writing
Mistral Large leads by +5.4
Lech Mazur Writing · evaluates creative writing ability, assessing prose quality, narrative coherence, and stylistic sophistication.
Mistral Large
69.0
Grok-2 (Dec 2024)
63.6
MATH level 5
Grok-2 (Dec 2024) leads by +39.1
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
Mistral Large
24.5
Grok-2 (Dec 2024)
63.5
OTIS Mock AIME 2024-2025
Grok-2 (Dec 2024) leads by +9.6
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Mistral Large
1.9
Grok-2 (Dec 2024)
11.4
SimpleBench
Grok-2 (Dec 2024) leads by +0.2
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
Mistral Large
7.0
Grok-2 (Dec 2024)
7.2
Full benchmark table
| Benchmark | Mistral Large | Grok-2 (Dec 2024) |
|---|---|---|
Aider · Code Editing | 60.2 | 58.6 |
FrontierMath-2025-02-28-Private FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning. | 0.3 | 0.7 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 18.4 | 38.4 |
Lech Mazur Writing Lech Mazur Writing · evaluates creative writing ability, assessing prose quality, narrative coherence, and stylistic sophistication. | 69.0 | 63.6 |
MATH level 5 MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics. | 24.5 | 63.5 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 1.9 | 11.4 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 7.0 | 7.2 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $2.00 | $6.00 | 128K tokens (~64 books) | $30.00 | |
| — | — | — | — |