Model comparison
Qwen3 14B vs Qwen3 8B
Qwen3 14B is the stronger model overall, scoring 35.0 to 33.3 on the Noometry Index. Qwen3 8B costs 2.0× less per token, which makes it the better buy when Qwen3 14B's lead doesn't matter for your workload.
Last verified . 11 shared benchmarks.
Alibaba (Qwen)Summary
- They share 11 benchmarks with published results for both. Qwen3 14B scores higher in 5 categories and Qwen3 8B in 1 category; 4 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3 14B leads 37.0 to 33.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 66.4% for Qwen3 14B and 56.1% for Qwen3 8B.
- Qwen3 8B is cheaper at $0.18 / $0.70 per million input/output tokens, against $0.35 / $1.40 for Qwen3 14B.
Side by side
| Qwen3 14B | Qwen3 8B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 35.0 | 33.3 |
| Released | 2025-04 | 2025-04 |
| Weights | Open | Open |
| Context window | 131K | 131K |
| Max output | 8K | 8K |
| Input $ / M tokens | $0.35 | $0.18 |
| Output $ / M tokens | $1.40 | $0.70 |
| Results tracked | 12 | 11 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 14B leads
Qwen3 14B: 36.7 (#191), Qwen3 8B: 33.4 (#245)
| Benchmark | Qwen3 14B | Qwen3 8B |
|---|---|---|
| SciCode | 31.6% | 22.6% |
Agentic & Tool Use Too close to call
Qwen3 14B: 29.4 (#78), Qwen3 8B: 29.9 (#71)
| Benchmark | Qwen3 14B | Qwen3 8B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 41% | 42.6% |
Reasoning Qwen3 14B leads
Qwen3 14B: 18.1 (#272), Qwen3 8B: 16.1 (#299)
| Benchmark | Qwen3 14B | Qwen3 8B |
|---|---|---|
| CritPt | 0% | 0% |
| Chess Puzzles | 4% | 5% |
| DTBench | 64% | 59.7% |
| LMCA | 18.2% | 8.8% |
| Epoch Capabilities Index | 138.23 | 136.17 |
| Kagi LLM Benchmark | 49.1% | — |
Math Qwen3 14B leads
Qwen3 14B: 37.0 (#102), Qwen3 8B: 33.3 (#176)
| Benchmark | Qwen3 14B | Qwen3 8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 56.1% |
Knowledge Qwen3 14B leads
Qwen3 14B: 39.2 (#124), Qwen3 8B: 36.1 (#164)
| Benchmark | Qwen3 14B | Qwen3 8B |
|---|---|---|
| GPQA Diamond | 63.8% | 56.8% |
| Vectara Hallucination Rate | 5.4% | 4.8% |
Long Context Too close to call
Qwen3 14B: 38.0 (#201), Qwen3 8B: 37.8 (#205)
| Benchmark | Qwen3 14B | Qwen3 8B |
|---|---|---|
| Fiction.LiveBench | 62.5% | 62.1% |
Frequently asked questions
Is Qwen3 14B better than Qwen3 8B?
Qwen3 14B is the stronger model overall, scoring 35.0 to 33.3 on the Noometry Index. Qwen3 8B costs 2.0× less per token, which makes it the better buy when Qwen3 14B's lead doesn't matter for your workload.
Which is cheaper, Qwen3 14B or Qwen3 8B?
Qwen3 8B is cheaper. It lists at $0.18 per million input tokens and $0.70 per million output tokens; Qwen3 14B lists at $0.35 and $1.40.
Is Qwen3 14B or Qwen3 8B better for coding?
Qwen3 14B scores higher on coding benchmarks: 36.7 versus 33.4 in the Noometry coding category.
Which has the bigger context window?
Both accept 131K tokens.
How many benchmarks do Qwen3 14B and Qwen3 8B share?
11 benchmarks have published results for both models. Qwen3 14B has 12 scored results on Noometry and Qwen3 8B has 11.