Model comparison
GPT-4 vs Qwen3-Next 80B-A3B Instruct
Qwen3-Next 80B-A3B Instruct is the stronger model overall, scoring 42.5 to 28.9 on the Noometry Index.
Last verified . 17 shared benchmarks.
Alibaba (Qwen)Summary
- They share 17 benchmarks with published results for both. GPT-4 scores higher in 1 category and Qwen3-Next 80B-A3B Instruct in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3-Next 80B-A3B Instruct leads 36.3 to 10.1.
- Qwen3-Next 80B-A3B Instruct is cheaper at $0.50 / $2 per million input/output tokens, against $30 / $60 for GPT-4.
- Qwen3-Next 80B-A3B Instruct accepts more context: 131K tokens versus 8K.
- Qwen3-Next 80B-A3B Instruct has downloadable open weights; the other is API-only.
Side by side
| GPT-4 | Qwen3-Next 80B-A3B Instruct | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 28.9 | 42.5 |
| Released | 2023-03-14 | 2025-09 |
| Weights | Proprietary | Open |
| Context window | 8K | 131K |
| Max output | 8K | 33K |
| Input $ / M tokens | $30 | $0.50 |
| Output $ / M tokens | $60 | $2 |
| Results tracked | 38 | 25 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3-Next 80B-A3B Instruct leads
GPT-4: 31.2 (#277), Qwen3-Next 80B-A3B Instruct: 41.9 (#101)
| Benchmark | GPT-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Coding | 1254 | 1441 |
| WeirdML | 12.4% | — |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| HumanEval+ | 79.3% | — |
Agentic & Tool Use Not comparable
GPT-4: —, Qwen3-Next 80B-A3B Instruct: —
| Benchmark | GPT-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| METR Time Horizons | 36.1% | — |
Reasoning Qwen3-Next 80B-A3B Instruct leads
GPT-4: 17.3 (#285), Qwen3-Next 80B-A3B Instruct: 30.1 (#79)
| Benchmark | GPT-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1428 |
| Kagi LLM Benchmark | — | 66.7% |
| Chess Puzzles | 4% | — |
| Mystery Game Puzzles | 12% | — |
| DTBench | 62.7% | — |
| LMCA | 17.1% | — |
| BIG-Bench Hard | 75.1% | — |
| Epoch Capabilities Index | 125.88 | — |
| ForecastBench | 57.8 | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |
Math Qwen3-Next 80B-A3B Instruct leads
GPT-4: 10.1 (#304), Qwen3-Next 80B-A3B Instruct: 36.3 (#118)
| Benchmark | GPT-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Math | 1269 | 1440 |
| OTIS Mock AIME 2024-2025 | 1.1% | — |
| Omni-MATH | — | 46.7% |
| MATH Level 5 | 23% | — |
| GSM8K | 92% | — |
Knowledge Qwen3-Next 80B-A3B Instruct leads
GPT-4: 18.4 (#277), Qwen3-Next 80B-A3B Instruct: 41.3 (#102)
| Benchmark | GPT-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Expert | 1211 | 1416 |
| GPQA Diamond | 35.7% | — |
| MMLU-Pro | — | 78.6% |
| Vectara Hallucination Rate | — | 9.3% |
| GPQA (HELM) | — | 63% |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |
Multilingual Qwen3-Next 80B-A3B Instruct leads
GPT-4: 40.8 (#210), Qwen3-Next 80B-A3B Instruct: 52.3 (#88)
| Benchmark | GPT-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Non-English | 1246 | 1407 |
| LMArena Chinese | 1242 | 1461 |
| LMArena French | 1282 | 1415 |
| LMArena German | 1251 | 1416 |
| LMArena Japanese | 1209 | 1398 |
| LMArena Korean | 1184 | 1366 |
| LMArena Russian | 1251 | 1403 |
| LMArena Spanish | 1261 | 1436 |
Instruction Following Qwen3-Next 80B-A3B Instruct leads
GPT-4: 65.5 (#217), Qwen3-Next 80B-A3B Instruct: 71.0 (#154)
| Benchmark | GPT-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Instruction Following | 1241 | 1388 |
| IFEval | — | 81% |
Long Context Too close to call
GPT-4: 37.7 (#207), Qwen3-Next 80B-A3B Instruct: 37.0 (#218)
| Benchmark | GPT-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Longer Query | 1244 | 1403 |
| Fiction.LiveBench | — | 55.6% |
Writing & Preference Qwen3-Next 80B-A3B Instruct leads
GPT-4: 35.0 (#263), Qwen3-Next 80B-A3B Instruct: 58.2 (#116)
| Benchmark | GPT-4 | Qwen3-Next 80B-A3B Instruct |
|---|---|---|
| LMArena Text | 1263 | 1417 |
| LMArena Creative Writing | 1244 | 1333 |
| LMArena Multi-Turn | 1257 | 1415 |
| EQ-Bench Creative Writing | 752 | — |
| WildBench | — | 80.7% |
Frequently asked questions
Is GPT-4 better than Qwen3-Next 80B-A3B Instruct?
Qwen3-Next 80B-A3B Instruct is the stronger model overall, scoring 42.5 to 28.9 on the Noometry Index.
Which is cheaper, GPT-4 or Qwen3-Next 80B-A3B Instruct?
Qwen3-Next 80B-A3B Instruct is cheaper. It lists at $0.50 per million input tokens and $2 per million output tokens; GPT-4 lists at $30 and $60.
Is GPT-4 or Qwen3-Next 80B-A3B Instruct better for coding?
Qwen3-Next 80B-A3B Instruct scores higher on coding benchmarks: 41.9 versus 31.2 in the Noometry coding category.
Which has the bigger context window?
Qwen3-Next 80B-A3B Instruct does, with 131K tokens against 8K.
How many benchmarks do GPT-4 and Qwen3-Next 80B-A3B Instruct share?
17 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Qwen3-Next 80B-A3B Instruct has 25.