Compare / Qwen3.5-4B vs gpt-oss-20b
Qwen3.5-4B vs gpt-oss-20b
Which is cheaper, which is more capable, and which does better at each kind of work, from 15 results on 2 sources that measured both models.
Of the 15 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.
Results as of 19 November 2026 (catalogue release 2026-11-19-3608133e8331). See also Alibaba vs OpenAI, every model from both labs.
Qwen3.5-4B or gpt-oss-20b? The short answer
Pick Qwen3.5-4B: it scores higher (Intelligence Index 13.1 against 10.0) at about the same price. Pick gpt-oss-20b only if speed matters more: it writes faster.
- Qwen3.5-4B scores higher on the Artificial Analysis Intelligence Index: 13.1 against 10.0.
- gpt-oss-20b writes faster: 277 tokens a second against 10.
- Qwen3.5-4B is newer: released 2 Mar 2026, against 5 Aug 2025 for gpt-oss-20b.
- On third-party benchmarks, on the results both models have: Qwen3.5-4B does better for reasoning and knowledge and writing.
Head to head
Every fact side by side
| Qwen3.5-4B | gpt-oss-20b | |
|---|---|---|
| Price per million tokens | Not on sale | $0.030 in · $0.14 out |
| Blended (3 in : 1 out) | – | $0.058 |
| Cheapest host, blended | – | $0.022 (PrimeIntellect, FP4) |
| Intelligence Index | 13.1 | 10.0 |
| Spring Prompt overall | – | – |
| Output speed, tokens a second | 10 | 277 |
| Context window | – | 131,072 tokens |
| Released | 2 Mar 2026 (newer) | 5 Aug 2025 |
| Weights | Closed | Open |
| Developer based in | China | the US |
Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.
Which is better for what
Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.
Every shared result
Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.
| Measure | Qwen3.5-4B | gpt-oss-20b |
|---|---|---|
| Artificial Analysis reported by Artificial Analysis · 3 results | ||
| Artificial Analysis Intelligence Indexindex score, higher is better 146 · 176 of 249 measured | 13.1thinking reasoning | 10.0low reasoning |
| GPQA Diamond% of questions, higher is better 128 · 159 of 232 measured | 77.1%thinking reasoning | 68.8%high reasoning |
| Humanity's Last Exam% of questions, higher is better 164 · 153 of 250 measured | 9.9%thinking reasoning | 11.0%high reasoning |
| UGI Leaderboard reported by UGI Leaderboard · 3 results | ||
| Writing score · UGIscore out of 100, higher is better 150 · 182 of 230 measured | 30.8think-prefill reasoning | 24.6low reasoning |
| Requested-length error · UGI% off the requested word count, lower is better 76 · 221 of 230 measured | 16.0%think-prefill reasoning | 98.0%low reasoning |
| Style adherence · UGIscore from 0 to 1, higher is better 175 · 187 of 230 measured | 0.32no reasoning | 0.31medium reasoning |
Show all 15 results (9 more)
| Measure | Qwen3.5-4B | gpt-oss-20b |
|---|---|---|
| Artificial Analysis reported by Artificial Analysis · 9 results | ||
| Terminal-Bench 2.1% of tasks, higher is better 90 · 97 of 131 measured | 25.8%thinking reasoning | 13.9%high reasoning |
| τ-bench banking% of tasks, higher is better 92 · 90 of 121 measured | 6.8%thinking reasoning | 7.0%high reasoning |
| Artificial Analysis Coding Indexindex score, higher is better 95 · 100 of 130 measured | 22.6thinking reasoning | 20.7high reasoning |
| IFBench% of instructions, higher is better 100 · 69 of 195 measured | 52.0%thinking reasoning | 65.1%high reasoning |
| Long-context reasoning (AA-LCR)% of questions, higher is better 143 · 184 of 241 measured | 63.0%thinking reasoning | 34.7%high reasoning |
| Output speedtokens per second, higher is better 78 · 6 of 78 measured | 9.71thinking reasoning | 277low reasoning |
| Terminal-Bench Hard% of tasks, higher is better 97 · 127 of 190 measured | 18.2%thinking reasoning | 10.6%high reasoning |
| Time to first answer tokenseconds, lower is better 16 · 44 of 78 measured | 0.70no reasoning | 7.62low reasoning |
| τ²-bench telecom% of tasks, higher is better 36 · 96 of 193 measured | 92.1%thinking reasoning | 60.2%high reasoning |
Compare them with other models
Quick answers
Is Qwen3.5-4B better than gpt-oss-20b?
Pick Qwen3.5-4B: it scores higher (Intelligence Index 13.1 against 10.0) at about the same price. Pick gpt-oss-20b only if speed matters more: it writes faster.
Which is faster, Qwen3.5-4B or gpt-oss-20b?
gpt-oss-20b writes about 277 tokens a second on its usual API, against 10 for Qwen3.5-4B (Artificial Analysis).
Which is better for agents and tool use, Qwen3.5-4B or gpt-oss-20b?
Qwen3.5-4B ranks 125 of 160 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 130 for gpt-oss-20b.
Which is better for reasoning and knowledge, Qwen3.5-4B or gpt-oss-20b?
Qwen3.5-4B ranks 124 of 188 models on sale for hard questions across science, maths and general knowledge, against 132 for gpt-oss-20b.
Which is better for writing, Qwen3.5-4B or gpt-oss-20b?
Qwen3.5-4B ranks 137 of 170 models on sale for what people prefer in blind comparisons, and judged writing quality, against 155 for gpt-oss-20b.
Run Qwen3.5-4B and gpt-oss-20b on your own prompt.
Benchmarks aren't your task. Test both side by side on your prompt, with every answer checked against your rules and the cost of each one.