Mistral Nemo Instruct 2407 — Benchmarks

Benchmark scores for Mistral Nemo Instruct 2407 aggregated from public leaderboards (5 boards), with how it ranks among open models and how it compares with its nearest neighbours. See hardware requirements for what you need to run it.

llmrun Composite Score →

65.4

90% range 58.2–72.7 · from 4 boards

CodingNot enough data yet
AgentsNot enough data yet
MathNot enough data yet
ScienceNot enough data yet
ReasoningNot enough data yet

Mistral Nemo Instruct 2407 reasoning benchmarks

Multi-step logic and problem-solving — can the model think through a hard problem rather than recall an answer.

GPQA Diamond
GPQA Diamond is a set of extremely hard, graduate-level science questions (physics, chemistry, biology) written by domain experts and filtered so that skilled non-experts with web access still fail. It measures genuine reasoning rather than memorization.
29.9%
29.9%
median 55.7%open leader 93.2%
Overall best: Gemini 4 Argon (high) 96.4%
16th percentile · #84 of 100 open · #288 of 309 all
#84 of 100
#288 of 309
Independent
Epoch AI · Jul 2024
BALROG
BALROG (Benchmarking Agentic LLM and VLM Reasoning On Games) tests long-horizon agentic reasoning across six game environments — BabyAI, Crafter, TextWorld, Baba Is AI, MiniHack and NetHack — scoring the average progress a model makes toward completing each game. It was built by UCL's DARK lab and collaborators.
17.6%
17.6%
median 16.2%open leader 34.9%
Overall best: GPT 6 Astra Max 68.3%
64th percentile · #6 of 15 open · #29 of 40 all
#6 of 15
#29 of 40
Independent
Epoch AI · Jul 2024
DTBench
DTBench is a set of handcrafted, expert-validated multiple-choice questions on decision theory — Newcomb-like problems and related puzzles — built by the Conceptual Reasoning Index project with Anthropic. It tests whether a model can follow an abstract argument rather than recall a fact.
48.6%
48.6%
median 68.7%open leader 93.9%
Overall best: Claude Opus 5.5 Max 98.9%
7th percentile · #65 of 70 open · #224 of 235 all
#65 of 70
#224 of 235
Lab-reported
Epoch AI · Jul 2024

Mistral Nemo Instruct 2407 math benchmarks

Symbolic and quantitative problem-solving, from competition math to multi-step calculation.

MATH Level 5
MATH Level 5 covers the hardest tier of competition-mathematics problems, testing multi-step symbolic and quantitative reasoning.
10.8%
10.8%
median 41.6%open leader 96.6%
Overall best: GPT 5 (Aug 07, 2025, high) 98.1%
16th percentile · #33 of 39 open · #100 of 107 all
#33 of 39
#100 of 107
Independent
Epoch AI · Jul 2024

Mistral Nemo Instruct 2407 knowledge benchmarks

Breadth and depth of factual knowledge across many subjects, paired with the reasoning to apply it.

MMLU-Pro
MMLU-Pro is a harder, cleaned-up successor to MMLU with ten answer choices and more reasoning-heavy questions across 14 subjects, measuring broad knowledge and reasoning together.
44.8%
44.8%
median 52.0%open leader 88.0%
Overall best: Gemini-3.1-Pro 91.2%
41st percentile · #87 of 146 open · #180 of 257 all
#87 of 146
#180 of 257
Independent
TIGER-Lab

Score vs VRAM in the 24 GB tier

Mistral Nemo Instruct 2407
Other models
Best Score for its VRAM
x: VRAM (GB) · y: llmrun Score

#32 of 36 that fit 24 GB

Each dot is a model at the best quant that fits a 24 GB GPU. The line steps up to the best Score available at each amount of VRAM; a dot below it is beaten by a model that needs no more memory.

Full 24 GB ranking →

Mistral Nemo Instruct 2407 vs its neighbours

Same boards for the nearest open models per llmrun Score and the best model that fits 24 GB. The highest score in each row is marked.

Board	Mistral Nemo Instruct 2407this model	Qwen3.8 27Bbest in classQ4_K_M · 18 GB	Gemma 2 9B ITnearest by ScoreQ4_K_M · 7.3 GB	Qwen2.5 7B Instructnearest by ScoreQ4_K_M · 5.8 GB	Llama 3.1 8B Instructnearest by ScoreQ4_K_M · 7.9 GB
llmrun Score	65.4	122.0	68.0	65.1	62.5
GPQA Diamond	29.9%	—	27.5%	35.5%	27.0%
BALROG	17.6%	—	—	7.8%	15.1%
DTBench	48.6%	88.0%	—	47.7%	50.9%
MATH Level 5	10.8%	—	21.0%	—	22.9%
MMLU-Pro	44.8%	—	52.1%	—	44.3%
Compare head to head		Compare	Compare	Compare	Compare

Scores aggregated from public benchmark sources (Epoch AI: 4 boards, TIGER-Lab: 1 board) — each linked from the benchmark pages. llmrun does not run these benchmarks. Percentiles and ranks are among open models in the llmrun catalogue; the median and leader marks refer to the same field.