Mistral Small 4 vs Qwen3 30B A3B Instruct 2507
Mistral Small 4 leads Qwen3 30B A3B Instruct 2507 on the Artificial Analysis Intelligence Index, 11.3 to 7.5, and Qwen3 30B A3B Instruct 2507 costs 2.0x less per output token ($0.30 vs $0.60 per million). Updated October 22, 2026; 10 benchmarks have a result for both.
Mistral Small 4 or Qwen3 30B A3B Instruct 2507: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick Mistral Small 4 for work measured by Humanity's Last Exam (Artificial Analysis run) or IFBench, where it leads.
- Shared results
- 10
- Mistral Small 4 only
- 5
- Qwen3 30B A3B Instruct 2507 only
- 35
- Like-for-like categories
- 5 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
Mistral Small 4
Mistral Small 4 leads, 11.3 against 7.5.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Mistral Small 4 has a AA Agentic Index result, 0.8.
Directional only
Long documents
Listed context window
No call
Both take 262K tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Qwen3 30B A3B Instruct 2507
Mistral Small 4 $0.0004, Qwen3 30B A3B Instruct 2507 $0.0002 per turn.
List prices
At a glance
- Coding index
- Mistral Small 4 not benchmarked, Qwen3 30B A3B Instruct 2507 not benchmarked.
- Widest category
- Agentic work, 13.2 points apart on average over 1 shared benchmark.
- Widest single gap
- τ²-Bench Telecom: Mistral Small 4 41.2% against Qwen3 30B A3B Instruct 2507 28.1%, 13.2 points apart on a 0 to 100 scale.
- Output price
- Mistral Small 4 $0.60 against Qwen3 30B A3B Instruct 2507 $0.30 per million output tokens, 2.0x apart.
- Context window
- Both take 262K tokens.
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 | Mistral Small 4 | Bars | Qwen3 30B A3B Instruct 2507 | Ties |
|---|---|---|---|---|
| Coding | 2 | 0 | ||
| Intelligence | 1 | 0 | ||
| Agentic work | 1 | 0 | ||
| Reasoning | 1 | 1 | ||
| General | 1 | 3 | ||
| All shared boards | 6 | 4 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index20
| Benchmark | Mistral Small 4 | Qwen3 30B A3B Instruct 2507 | Leader |
|---|---|---|---|
| AA Coding Index | 26.6 | Not measured | Only Mistral Small 4 measured |
| Vibe Code Bench (Vals AI) | 0.0% | Not measured | Only Mistral Small 4 measured |
| LiveCodeBench (Artificial Analysis run) | Not measured | 70.7% | Only Qwen3 30B A3B Instruct 2507 measured |
| SciCode | 38.8% | 33.0% | Mistral Small 4 |
| Terminal-Bench Hard (Artificial Analysis run) | 17.4% | 6.1% | Mistral Small 4 |
| ALE-Bench | 497.6 | Not measured | Only Mistral Small 4 measured |
| LMArena Text: coding | Not measured | 1416 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: software and IT | Not measured | 1417 | Only Qwen3 30B A3B Instruct 2507 measured |
| Neither model has a result on the other 35 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Mistral Small 4 | Qwen3 30B A3B Instruct 2507 | Leader |
|---|---|---|---|
| AA Intelligence Index | 11.3 | 7.5 | Mistral Small 4 |
| Neither model has a result on the other 1 board here | |||
Agentic workLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Mistral Small 4 | Qwen3 30B A3B Instruct 2507 | Leader |
|---|---|---|---|
| AA Agentic Index | 0.8 | Not measured | Only Mistral Small 4 measured |
| τ²-Bench Telecom | 41.2% | 28.1% | Mistral Small 4 |
| Berkeley Function Calling Leaderboard (BFCL) | Not measured | 41.4% | Only Qwen3 30B A3B Instruct 2507 measured |
| Neither model has a result on the other 39 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | Mistral Small 4 | Qwen3 30B A3B Instruct 2507 | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | 76.9% | 70.7% | Mistral Small 4 |
| Humanity's Last Exam (Artificial Analysis run) | 9.9% | 10.3% | Qwen3 30B A3B Instruct 2507 |
| GPQA Diamond (Epoch AI run) | Not measured | 55.6% | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: hard prompts | Not measured | 1398 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: expert | Not measured | 1397 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: hard prompts in English | Not measured | 1406 | Only Qwen3 30B A3B Instruct 2507 measured |
| Kagi LLM Benchmark | 40.5% | Not measured | Only Mistral Small 4 measured |
| Neither model has a result on the other 16 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index5 boards, none shared
| Benchmark | Mistral Small 4 | Qwen3 30B A3B Instruct 2507 | Leader |
|---|---|---|---|
| AA Math Index | Not measured | 66.3 | Only Qwen3 30B A3B Instruct 2507 measured |
| AIME 2025 (Artificial Analysis run) | Not measured | 66.3% | Only Qwen3 30B A3B Instruct 2507 measured |
| OTIS Mock AIME | Not measured | 62.2% | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: math | Not measured | 1394 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: mathematical work | Not measured | 1411 | Only Qwen3 30B A3B Instruct 2507 measured |
| Neither model has a result on the other 9 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index5 boards, none shared
| Benchmark | Mistral Small 4 | Qwen3 30B A3B Instruct 2507 | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | Not measured | 80.5% | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: science | Not measured | 1393 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: medicine and healthcare | Not measured | 1392 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: legal and government | Not measured | 1387 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: business and finance | Not measured | 1389 | Only Qwen3 30B A3B Instruct 2507 measured |
| Neither model has a result on the other 15 boards here | |||
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Mistral Small 4 | Qwen3 30B A3B Instruct 2507 | Leader |
|---|---|---|---|
| Neither model has a result on 9 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index13
| Benchmark | Mistral Small 4 | Qwen3 30B A3B Instruct 2507 | Leader |
|---|---|---|---|
| IFBench | 48.2% | 50.7% | Qwen3 30B A3B Instruct 2507 |
| Long context reasoning (AA-LCR) | 49.7% | 61.3% | Qwen3 30B A3B Instruct 2507 |
| DTBench | 70.9% | 67.2% | Mistral Small 4 |
| LMCA | 20.6% | 22.4% | Qwen3 30B A3B Instruct 2507 |
| LMArena Text | Not measured | 1384 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: instruction following | Not measured | 1363 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: creative writing | Not measured | 1317 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: English | Not measured | 1391 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: non-English | Not measured | 1372 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: Chinese | Not measured | 1432 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: German | Not measured | 1379 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: Japanese | Not measured | 1337 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: Korean | Not measured | 1330 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: French | Not measured | 1419 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: Spanish | Not measured | 1404 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: Russian | Not measured | 1369 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: Polish | Not measured | 1353 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: multi-turn | Not measured | 1378 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: longer queries | Not measured | 1378 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: writing and language | Not measured | 1336 | Only Qwen3 30B A3B Instruct 2507 measured |
| LMArena Text: entertainment and media | Not measured | 1334 | Only Qwen3 30B A3B Instruct 2507 measured |
| Neither model has a result on the other 27 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Mistral Small 4 | Qwen3 30B A3B Instruct 2507 | 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
Qwen3 30B A3B Instruct 2507 costs less
Repository review
50K input + 3K output tokens
Qwen3 30B A3B Instruct 2507 costs less
Agent loop
200K input + 10K output tokens
Qwen3 30B A3B Instruct 2507 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 Small 4 | Qwen3 30B A3B Instruct 2507 |
|---|---|---|
| Maker | Mistral AI | Qwen |
| Weights | Open weights | Open weights |
| Released | March 16, 2026 | July 29, 2025 |
| Context window | 262K | 262K |
| Input $/Mtok | $0.15 | $0.09 |
| Output $/Mtok | $0.60 | $0.30 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 1 | 0 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Mistral Small 4 better than Qwen3 30B A3B Instruct 2507 for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Mistral Small 4 for work measured by Humanity's Last Exam (Artificial Analysis run) or IFBench, where it leads.
Which is cheaper, Mistral Small 4 or Qwen3 30B A3B Instruct 2507?
Mistral Small 4 costs $0.15 per million input tokens and $0.60 per million output tokens; Qwen3 30B A3B Instruct 2507 costs $0.09 and $0.30. Cost per solved task is known for neither of them.
Which has the bigger context window, Mistral Small 4 or Qwen3 30B A3B Instruct 2507?
Neither: both take 262K tokens.
How many benchmarks is the Mistral Small 4 vs Qwen3 30B A3B Instruct 2507 comparison based on?
10 benchmarks have a result for both models as of October 22, 2026. Mistral Small 4's coding index rests on 1 independent benchmark and Qwen3 30B A3B Instruct 2507's on 0. Where only one model has a result, the page says so instead of naming a winner.