Llama 3.3 70B Instruct vs Qwen3.5-27B
Qwen3.5-27B leads Llama 3.3 70B Instruct on the Artificial Analysis Intelligence Index, 22.9 to 7.7, and Llama 3.3 70B Instruct costs 2.2x less per output token ($0.72 vs $1.56 per million). Updated October 22, 2026; 38 benchmarks have a result for both.
Llama 3.3 70B Instruct or Qwen3.5-27B: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick Llama 3.3 70B Instruct for output tokens 2.2x cheaper ($0.72 vs $1.56 per million) and work measured by Terminal-Bench Hard (Artificial Analysis run) or WeirdML, where it leads.
- Shared results
- 38
- Llama 3.3 70B Instruct only
- 14
- Qwen3.5-27B only
- 15
- Like-for-like categories
- 7 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
Qwen3.5-27B
Qwen3.5-27B leads, 22.9 against 7.7.
Same board, both measured
Long documents
Listed context window
Qwen3.5-27B
Qwen3.5-27B takes 262K tokens against 131K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Llama 3.3 70B Instruct
Llama 3.3 70B Instruct $0.0008, Qwen3.5-27B $0.0010 per turn.
List prices
At a glance
- Coding index
- Llama 3.3 70B Instruct not benchmarked, Qwen3.5-27B not benchmarked.
- Widest category
- Agentic work, 67.3 points apart on average over 1 shared benchmark.
- Widest single gap
- τ²-Bench Telecom: Llama 3.3 70B Instruct 26.6% against Qwen3.5-27B 93.9%, 67.3 points apart on a 0 to 100 scale.
- Output price
- Llama 3.3 70B Instruct $0.72 against Qwen3.5-27B $1.56 per million output tokens, 2.2x apart.
- Context window
- Llama 3.3 70B Instruct takes 131K tokens, Qwen3.5-27B 262K.
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 | Llama 3.3 70B Instruct | Bars | Qwen3.5-27B | Ties |
|---|---|---|---|---|
| Coding | 0 | 4 | ||
| Intelligence | 0 | 1 | ||
| Agentic work | 0 | 1 | ||
| Reasoning | 0 | 5 | ||
| Math | 0 | 2 | ||
| Knowledge | 0 | 4 | ||
| General | 0 | 21 | ||
| All shared boards | 0 | 38 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index04
| Benchmark | Llama 3.3 70B Instruct | Qwen3.5-27B | Leader |
|---|---|---|---|
| LMArena WebDev | Not measured | 1358 | Only Qwen3.5-27B measured |
| AA Coding Index | 11.9 | Not measured | Only Llama 3.3 70B Instruct measured |
| LiveCodeBench (Artificial Analysis run) | 28.8% | Not measured | Only Llama 3.3 70B Instruct measured |
| Terminal-Bench Hard (Artificial Analysis run) | 3.0% | 32.6% | Qwen3.5-27B |
| WeirdML | 14.4% | 39.5% | Qwen3.5-27B |
| ALE-Bench | Not measured | 349.4 | Only Qwen3.5-27B measured |
| LMArena Text: coding | 1268 | 1427 | Qwen3.5-27B |
| LMArena Text: software and IT | 1273 | 1431 | Qwen3.5-27B |
| LMArena WebDev: React | Not measured | 1346 | Only Qwen3.5-27B measured |
| LMArena WebDev: HTML | Not measured | 1386 | Only Qwen3.5-27B measured |
| Neither model has a result on the other 29 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Llama 3.3 70B Instruct | Qwen3.5-27B | Leader |
|---|---|---|---|
| AA Intelligence Index | 7.7 | 22.9 | Qwen3.5-27B |
| Epoch Capabilities Index | 127.3 | Not measured | Only Llama 3.3 70B Instruct measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Llama 3.3 70B Instruct | Qwen3.5-27B | Leader |
|---|---|---|---|
| τ²-Bench Telecom | 26.6% | 93.9% | Qwen3.5-27B |
| BALROG | 23.0% | Not measured | Only Llama 3.3 70B Instruct measured |
| Vending-Bench 2 | Not measured | $202 | Only Qwen3.5-27B measured |
| Berkeley Function Calling Leaderboard (BFCL) | 31.9% | Not measured | Only Llama 3.3 70B Instruct measured |
| Neither model has a result on the other 37 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index05
| Benchmark | Llama 3.3 70B Instruct | Qwen3.5-27B | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | 49.8% | 85.8% | Qwen3.5-27B |
| Humanity's Last Exam (Artificial Analysis run) | 3.6% | 23.9% | Qwen3.5-27B |
| GPQA Diamond (Epoch AI run) | 47.4% | Not measured | Only Llama 3.3 70B Instruct measured |
| SimpleBench | 19.9% | Not measured | Only Llama 3.3 70B Instruct measured |
| LMArena Text: hard prompts | 1257 | 1415 | Qwen3.5-27B |
| LMArena Text: expert | 1225 | 1427 | Qwen3.5-27B |
| LMArena Text: hard prompts in English | 1277 | 1429 | Qwen3.5-27B |
| LMArena Vision: diagrams | Not measured | 1249 | Only Qwen3.5-27B measured |
| LMArena Vision: homework | Not measured | 1256 | Only Qwen3.5-27B measured |
| Neither model has a result on the other 13 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | Llama 3.3 70B Instruct | Qwen3.5-27B | Leader |
|---|---|---|---|
| AA Math Index | 7.7 | Not measured | Only Llama 3.3 70B Instruct measured |
| AIME 2025 (Artificial Analysis run) | 7.7% | Not measured | Only Llama 3.3 70B Instruct measured |
| OTIS Mock AIME | 5.1% | Not measured | Only Llama 3.3 70B Instruct measured |
| MATH level 5 | 41.6% | Not measured | Only Llama 3.3 70B Instruct measured |
| LMArena Text: math | 1267 | 1429 | Qwen3.5-27B |
| LMArena Text: mathematical work | 1267 | 1429 | Qwen3.5-27B |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index04
| Benchmark | Llama 3.3 70B Instruct | Qwen3.5-27B | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | 71.3% | Not measured | Only Llama 3.3 70B Instruct measured |
| MMLU | 86.3% | Not measured | Only Llama 3.3 70B Instruct measured |
| LMArena Text: science | 1284 | 1437 | Qwen3.5-27B |
| LMArena Text: medicine and healthcare | 1275 | 1425 | Qwen3.5-27B |
| LMArena Text: legal and government | 1284 | 1405 | Qwen3.5-27B |
| LMArena Text: business and finance | 1243 | 1399 | Qwen3.5-27B |
| Neither model has a result on the other 12 boards here | |||
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Llama 3.3 70B Instruct | Qwen3.5-27B | Leader |
|---|---|---|---|
| Neither model has a result on 9 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index021
| Benchmark | Llama 3.3 70B Instruct | Qwen3.5-27B | Leader |
|---|---|---|---|
| IFBench | 47.1% | 75.6% | Qwen3.5-27B |
| Long context reasoning (AA-LCR) | 15.7% | 77.7% | Qwen3.5-27B |
| DTBench | 59.5% | 82.4% | Qwen3.5-27B |
| LMCA | 17.5% | 34.0% | Qwen3.5-27B |
| Fiction.LiveBench | 33.3% | Not measured | Only Llama 3.3 70B Instruct measured |
| LMArena Text | 1274 | 1409 | Qwen3.5-27B |
| LMArena Text: instruction following | 1242 | 1393 | Qwen3.5-27B |
| LMArena Text: creative writing | 1250 | 1362 | Qwen3.5-27B |
| LMArena Vision | Not measured | 1241 | Only Qwen3.5-27B measured |
| LMArena Text: English | 1303 | 1425 | Qwen3.5-27B |
| LMArena Text: non-English | 1236 | 1390 | Qwen3.5-27B |
| LMArena Text: Chinese | 1217 | 1478 | Qwen3.5-27B |
| LMArena Text: German | 1251 | 1392 | Qwen3.5-27B |
| LMArena Text: Japanese | 1150 | 1346 | Qwen3.5-27B |
| LMArena Text: Korean | 1142 | 1359 | Qwen3.5-27B |
| LMArena Text: French | 1282 | 1410 | Qwen3.5-27B |
| LMArena Text: Spanish | 1269 | 1407 | Qwen3.5-27B |
| LMArena Text: Russian | 1252 | 1390 | Qwen3.5-27B |
| LMArena Text: Polish | 1259 | 1375 | Qwen3.5-27B |
| LMArena Text: multi-turn | 1280 | 1410 | Qwen3.5-27B |
| LMArena Text: longer queries | 1255 | 1413 | Qwen3.5-27B |
| LMArena Text: writing and language | 1255 | 1380 | Qwen3.5-27B |
| LMArena Text: entertainment and media | 1241 | 1363 | Qwen3.5-27B |
| LMArena Vision: English | Not measured | 1246 | Only Qwen3.5-27B measured |
| LMArena Vision: Chinese | Not measured | 1301 | Only Qwen3.5-27B measured |
| LMArena Vision: reading text | Not measured | 1246 | Only Qwen3.5-27B measured |
| LMArena Vision: humor | Not measured | 1227 | Only Qwen3.5-27B measured |
| LMArena Vision: captioning | Not measured | 1170 | Only Qwen3.5-27B measured |
| LMArena Vision: recognizing things | Not measured | 1210 | Only Qwen3.5-27B measured |
| LMArena Vision: creative writing | Not measured | 1224 | Only Qwen3.5-27B measured |
| Neither model has a result on the other 16 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Llama 3.3 70B Instruct | Qwen3.5-27B | 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
Llama 3.3 70B Instruct costs less
Repository review
50K input + 3K output tokens
Qwen3.5-27B costs less
Agent loop
200K input + 10K output tokens
Qwen3.5-27B 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 | Meta Llama 3.3 70B Instruct | Qwen3.5-27B |
|---|---|---|
| Maker | Meta Llama | Qwen |
| Weights | Open weights | Open weights |
| Released | December 6, 2024 | February 25, 2026 |
| Context window | 131K | 262K |
| Input $/Mtok | $0.45 | $0.195 |
| Output $/Mtok | $0.72 | $1.56 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 0 | 1 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Llama 3.3 70B Instruct better than Qwen3.5-27B for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Llama 3.3 70B Instruct for output tokens 2.2x cheaper ($0.72 vs $1.56 per million) and work measured by Terminal-Bench Hard (Artificial Analysis run) or WeirdML, where it leads.
Which is cheaper, Llama 3.3 70B Instruct or Qwen3.5-27B?
Llama 3.3 70B Instruct costs $0.45 per million input tokens and $0.72 per million output tokens; Qwen3.5-27B costs $0.195 and $1.56. Cost per solved task is known for neither of them.
Which has the bigger context window, Llama 3.3 70B Instruct or Qwen3.5-27B?
Qwen3.5-27B, with 262K tokens against 131K.
How many benchmarks is the Llama 3.3 70B Instruct vs Qwen3.5-27B comparison based on?
38 benchmarks have a result for both models as of October 22, 2026. Llama 3.3 70B Instruct's coding index rests on 0 independent benchmarks and Qwen3.5-27B's on 1. Where only one model has a result, the page says so instead of naming a winner.