Noometry

Mistral AI, open weights

Mixtral 8x7B

Mixtral 8x7B by Mistral AI ranks 334th of 354 ranked models on the Noometry Index as of October 2026, with a score of 27.1. Its strongest category is long context, where it ranks 260th. API pricing starts at $0.70 per million input tokens and $0.70 per million output tokens, with a 32K-token context window.

Last verified

Noometry rank
#334 of 354
Index score
27.1
Evidence
Confirmed 38 results
Provider
Mistral AI
Released
December 11, 2023
Weights
Open weights
Reasoning
No
Context window
32K
Max output
32K
Input price
$0.70 / M
Output price
$0.70 / M
Blended price
$0.70 / M
Output speed
Not measured
Value
#121 of 219
Knowledge cutoff
January 2024
Input
text

Category scores

Each category score combines every public result we have in that category.

  1. Coding 32.8
  2. Reasoning 18.2
  3. Math 18.8
  4. Knowledge 11.0
  5. Multilingual 29.6
  6. Instruction Following 51.0
  7. Long Context 33.4
  8. Writing & Preference 34.2
CategoryScoreRankResults
Coding32.8#2691
Reasoning18.2#2852
Math18.8#2893
Knowledge11.0#3014
Multilingual29.6#2661
Instruction Following51.0#2972
Long Context33.4#2601
Writing & Preference34.2#2704

Strengths and weaknesses

Categories where Mixtral 8x7B places highest and lowest among the models ranked in each, with its score against that category's median.

Strongest categories

CategoryScorevs medianRank
Coding32.8−6.0#269 of 340, top 80%
Reasoning18.2−5.4#285 of 350, top 82%
Writing & Preference34.2−19.6#270 of 312, top 87%

Weakest categories

CategoryScorevs medianRank
Instruction Following51.0−20.3#297 of 305, top 98%
Knowledge11.0−26.3#301 of 314, top 96%
Multilingual29.6−17.8#266 of 297, top 90%

Closest competitors

The models ranked just above and below Mixtral 8x7B. When scores are this close, price and speed are often the better way to choose.

ModelRankScoreBlended $/MSpeed
Llama 4 Scout#33027.7$0.15272Compare
Llama 3.2 90B#33127.5——Compare
Gemini 1.0 Pro#33227.3——Compare
Mixtral 8x22B#33327.1$3—Compare
Qwen Turbo#33527.1$0.0875—Compare
Qwen3-1.7B#33626.6——Compare
Mistral Nemo#33726.4$0.15—Compare
Ministral 3B#33826.2$0.10—Compare

Sponsored placements are available on pages like this one. Advertise on Noometry

Benchmark results

Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.

Coding

BenchmarkScorePositionSettingSourceDate
LMArena Coding1126#263 of 294, top 90%LMArena2026-10-08
HumanEval+39.6%#37 of 45, top 83%EvalPlus
MBPP+49.7%#33 of 38, top 87%EvalPlus

Reasoning

BenchmarkScorePositionSettingSourceDate
LMArena Hard Prompts1115#261 of 297, top 88%LMArena2026-10-08
DTBench49.6%#138 of 151, top 92%Epoch AI
Adversarial NLI55.2%#5 of 9, top 56%Epoch AI
Epoch Capabilities Index118.47#176 of 213, top 83%Epoch AI2023-12-11
ForecastBench56.3#65 of 72, top 91%Epoch AI
HellaSwag86.7%#7 of 29, top 25%Epoch AI
PIQA83.6%#9 of 27, top 34%Epoch AI
WinoGrande77.2%#19 of 43, top 45%Epoch AI

Math

BenchmarkScorePositionSettingSourceDate
Omni-MATH10.5%#56 of 57, top 99%HELM Capabilities
LMArena Math1147#253 of 285, top 89%LMArena2026-10-08
MATH Level 510%#74 of 79, top 94%Epoch AI2025-01-27
MATH Level 59.3%Epoch AI2025-01-27
GSM8K74.4%#14 of 38, top 37%Epoch AI

Knowledge

BenchmarkScorePositionSettingSourceDate
GPQA Diamond29.8%Epoch AI2025-01-27
GPQA Diamond30.6%#169 of 186, top 91%Epoch AI2025-01-27
MMLU-Pro33.5%#56 of 58, top 97%HELM Capabilities
GPQA (HELM)29.6%#55 of 57, top 97%HELM Capabilities
LMArena Expert1088#253 of 273, top 93%LMArena2026-10-08
ARC (AI2) Challenge87.3%#8 of 39, top 21%Epoch AI
MMLU70.6%#44 of 81, top 55%Epoch AI
OpenBookQA85.8%#5 of 19, top 27%Epoch AI
TriviaQA82.2%#9 of 25, top 36%Epoch AI

Multilingual

BenchmarkScorePositionSettingSourceDate
LMArena Non-English1077#266 of 297, top 90%LMArena2026-10-08
LMArena Chinese1055#266 of 285, top 94%LMArena2026-10-08
LMArena French1166#200 of 223, top 90%LMArena2026-10-08
LMArena German1114#208 of 231, top 91%LMArena2026-10-08
LMArena Japanese931#207 of 211, top 99%LMArena2026-10-08
LMArena Korean968#204 of 213, top 96%LMArena2026-10-08
LMArena Russian1090#260 of 283, top 92%LMArena2026-10-08
LMArena Spanish1111#214 of 226, top 95%LMArena2026-10-08

Instruction Following

BenchmarkScorePositionSettingSourceDate
IFEval57.5%#56 of 57, top 99%HELM Capabilities
LMArena Instruction Following1109#266 of 298, top 90%LMArena2026-10-08

Long Context

BenchmarkScorePositionSettingSourceDate
LMArena Longer Query1103#266 of 291, top 92%LMArena2026-10-08

Writing & Preference

BenchmarkScorePositionSettingSourceDate
LMArena Text1132#262 of 297, top 89%LMArena2026-10-08
LMArena Creative Writing1109#261 of 295, top 89%LMArena2026-10-08
WildBench67.3%#54 of 57, top 95%HELM Capabilities
LMArena Multi-Turn1115#260 of 295, top 89%LMArena2026-10-08

API pricing by provider

RouteInput $/MOutput $/MCached input $/MChecked
mistral$0.70$0.70—2026-10-10

All Mistral AI API prices → Estimate your cost →

Compare Mixtral 8x7B

Other Mistral AI models

Frequently asked questions

How good is Mixtral 8x7B?

Mixtral 8x7B by Mistral AI ranks 334th of 354 ranked models on the Noometry Index as of October 2026, with a score of 27.1. Its strongest category is long context, where it ranks 260th. API pricing starts at $0.70 per million input tokens and $0.70 per million output tokens, with a 32K-token context window.

How much does Mixtral 8x7B cost?

Mixtral 8x7B costs $0.70 per million input tokens and $0.70 per million output tokens on Mistral AI's own API.

What is Mixtral 8x7B's context window?

Mixtral 8x7B accepts up to 32K tokens of input and can write up to 32K tokens in one response.

Is Mixtral 8x7B open source?

Yes. Mixtral 8x7B's weights are downloadable; check the license for commercial terms.

What are Mixtral 8x7B's strengths and weaknesses?

Relative to other ranked models, Mixtral 8x7B places best in coding, reasoning, writing & preference and lowest in instruction following, knowledge, multilingual.

What is Mixtral 8x7B best at?

Its best category is long context, where it ranks 260th on Noometry.