Ministral 3 14B 2512 vs Mixtral 8x22B Instruct
Ministral 3 14B 2512 leads Mixtral 8x22B Instruct on the Artificial Analysis Intelligence Index, 6.0 to 5.7, and Ministral 3 14B 2512 costs 30.0x less per output token ($0.20 vs $6.00 per million). Updated October 22, 2026; 1 benchmarks have a result for both.
Ministral 3 14B 2512 or Mixtral 8x22B Instruct: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick Ministral 3 14B 2512 for output tokens 30.0x cheaper ($0.20 vs $6.00 per million) and a larger context window (262K against 66K tokens).
- Shared results
- 1
- Ministral 3 14B 2512 only
- 6
- Mixtral 8x22B Instruct only
- 7
- Like-for-like categories
- 1 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.
Coding
Vibe Coding Index, six public coding benchmarks
No call
Neither has a coding score yet.
One side not measured
Intelligence
AA Intelligence Index
Ministral 3 14B 2512
Ministral 3 14B 2512 leads, 6.0 against 5.7.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Ministral 3 14B 2512 has a AA Agentic Index result, 1.1.
Directional only
Web design
Design Arena: websites
No call
Only Ministral 3 14B 2512 has a Design Arena: websites result, 1081.
Directional only
Long documents
Listed context window
Ministral 3 14B 2512
Ministral 3 14B 2512 takes 262K tokens against 66K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Ministral 3 14B 2512
Ministral 3 14B 2512 $0.0003, Mixtral 8x22B Instruct $0.0050 per turn.
List prices
At a glance
- Coding index
- Ministral 3 14B 2512 not benchmarked, Mixtral 8x22B Instruct not benchmarked.
- Widest category
- Intelligence, 0.3 points apart on average over 1 shared benchmark.
- Widest single gap
- AA Intelligence Index: Ministral 3 14B 2512 6.0 against Mixtral 8x22B Instruct 5.7, 0.3 points apart on a 0 to 100 scale.
- Output price
- Ministral 3 14B 2512 $0.20 against Mixtral 8x22B Instruct $6.00 per million output tokens, 30.0x apart.
- Context window
- Ministral 3 14B 2512 takes 262K tokens, Mixtral 8x22B Instruct 66K.
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.
| Category | Ministral 3 14B 2512 | Bars | Mixtral 8x22B Instruct | Ties |
|---|---|---|---|---|
| Intelligence | 1 | 0 | ||
| All shared boards | 1 | 0 |
CodingDirectional onlySix of these boards feed the Vibe Coding Index3 boards, none shared
| Benchmark | Ministral 3 14B 2512 | Mixtral 8x22B Instruct | Leader |
|---|---|---|---|
| AA Coding Index | 14.4 | Not measured | Only Ministral 3 14B 2512 measured |
| LiveCodeBench (Artificial Analysis run) | Not measured | 14.8% | Only Mixtral 8x22B Instruct measured |
| SciCode | Not measured | 0.0% | Only Mixtral 8x22B Instruct measured |
| Neither model has a result on the other 36 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Ministral 3 14B 2512 | Mixtral 8x22B Instruct | Leader |
|---|---|---|---|
| AA Intelligence Index | 6.0 | 5.7 | Ministral 3 14B 2512 |
| Epoch Capabilities Index | Not measured | 122.0 | Only Mixtral 8x22B Instruct measured |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
| Benchmark | Ministral 3 14B 2512 | Mixtral 8x22B Instruct | Leader |
|---|---|---|---|
| AA Agentic Index | 1.1 | Not measured | Only Ministral 3 14B 2512 measured |
| Neither model has a result on the other 40 boards here | |||
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index2 boards, none shared
| Benchmark | Ministral 3 14B 2512 | Mixtral 8x22B Instruct | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | Not measured | 33.2% | Only Mixtral 8x22B Instruct measured |
| Humanity's Last Exam (Artificial Analysis run) | Not measured | 4.0% | Only Mixtral 8x22B Instruct measured |
| Neither model has a result on the other 20 boards here | |||
MathNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Ministral 3 14B 2512 | Mixtral 8x22B Instruct | Leader |
|---|---|---|---|
| Neither model has a result on 13 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
| Benchmark | Ministral 3 14B 2512 | Mixtral 8x22B Instruct | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | Not measured | 53.7% | Only Mixtral 8x22B Instruct measured |
| Neither model has a result on the other 17 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index4 boards, none shared
| Benchmark | Ministral 3 14B 2512 | Mixtral 8x22B Instruct | Leader |
|---|---|---|---|
| Design Arena: websites | 1081 | Not measured | Only Ministral 3 14B 2512 measured |
| Design Arena: game dev | 1037 | Not measured | Only Ministral 3 14B 2512 measured |
| Design Arena: 3D | 989 | Not measured | Only Ministral 3 14B 2512 measured |
| Design Arena: all code categories | 1064 | Not measured | Only Ministral 3 14B 2512 measured |
| Neither model has a result on the other 5 boards here | |||
GeneralDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
| Benchmark | Ministral 3 14B 2512 | Mixtral 8x22B Instruct | Leader |
|---|---|---|---|
| DTBench | Not measured | 55.1% | Only Mixtral 8x22B Instruct measured |
| Neither model has a result on the other 45 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Ministral 3 14B 2512 | Mixtral 8x22B Instruct | Leader |
|---|---|---|---|
| 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
Ministral 3 14B 2512 costs less
Repository review
50K input + 3K output tokens
Ministral 3 14B 2512 costs less
Agent loop
200K input + 10K output tokens
Ministral 3 14B 2512 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.
| Specification | Mistral Ministral 3 14B 2512 | Mistral Mixtral 8x22B Instruct |
|---|---|---|
| Maker | Mistral AI | Mistral AI |
| Weights | Open weights | Open weights |
| Released | December 2, 2025 | April 17, 2024 |
| Context window | 262K | 66K |
| Input $/Mtok | $0.20 | $2.00 |
| Output $/Mtok | $0.20 | $6.00 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 0 | 0 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Ministral 3 14B 2512 better than Mixtral 8x22B Instruct for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Ministral 3 14B 2512 for output tokens 30.0x cheaper ($0.20 vs $6.00 per million) and a larger context window (262K against 66K tokens).
Which is cheaper, Ministral 3 14B 2512 or Mixtral 8x22B Instruct?
Ministral 3 14B 2512 costs $0.20 per million input tokens and $0.20 per million output tokens; Mixtral 8x22B Instruct costs $2.00 and $6.00. Cost per solved task is known for neither of them.
Which has the bigger context window, Ministral 3 14B 2512 or Mixtral 8x22B Instruct?
Ministral 3 14B 2512, with 262K tokens against 66K.
How many benchmarks is the Ministral 3 14B 2512 vs Mixtral 8x22B Instruct comparison based on?
1 benchmarks have a result for both models as of October 22, 2026. Ministral 3 14B 2512's coding index rests on 0 independent benchmarks and Mixtral 8x22B Instruct's on 0. Where only one model has a result, the page says so instead of naming a winner.