AlibabaQwen2.5 Coder 32B Instruct
Qwen2.5 Coder 32B Instruct is an Alibaba open-weights model, released 18 Sept 2024. Its best configuration ranks #609 of 785 on the BenchLeader Index at 42.1 ±7.6, in the lower half. The ± is the point: 386 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in reasoning (44) and lowest in coding (27). At $0.745 per million tokens blended it is mid-priced. Output speed of 30 tokens per second puts it in the slowest quarter, with a first token in 1.6 s. Last measured 22 Oct 2026.
- Blended price
- $0.745/M
- $0.660 in · $1.00 out
- Output speed
- 30 tok/s
- OpenRouter traffic, 7-day median; not yet measured by Artificial Analysis
- First answer
- 1.62 s
- Context
- 131k
- Full answer
- –
- Cost per run
- –
- Released
- 18 Sept 2024
- Overall index42
- Reasoning44
- Coding27
- Maths43
- Human preference42
- Composite37
How its index has moved
8 Sept 2026 to 21 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Hard Prompts | 1303#262 | Arena (LMArena) | |
| GPQA Diamond (AA)not in index | 41.7%#470 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 3.5%#592 | Artificial Analysis | |
| HellaSwagnot in index | 83.0%#16 | Epoch AI Benchmarking Hub | |
| WinoGrandenot in index | 80.8%#16 | Epoch AI Benchmarking Hub | |
| ARC (AI2 Challenge)not in index | 70.5%#23 | Epoch AI Benchmarking Hub |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Coding | 1342#252 | Arena (LMArena) | |
| Aider Polyglot | 16.4%#41 | Aider polyglot leaderboard | |
| SWE-bench Verified (bash only) | 9.0%#45 | SWE-bench | |
| SWE-bench Verified (any scaffold)not in index | 47.0%#38 | SWE-bench |
Maths
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Maths | 1269#256 | Arena (LMArena) | |
| GSM8Knot in index | 93.0%#3 | Epoch AI Benchmarking Hub |
Knowledge
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| MMLUnot in index | 79.1%#25 | Epoch AI Benchmarking Hub |
Instruction following
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Instruction Followingnot in index | 1264#270 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1271#279 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1207#289 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1251#276 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1287#267 | Arena (LMArena) |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Epoch Capabilities Indexnot in index | 119.5#176 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2 | 6.7#508 | Artificial Analysis |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Cloudflare | 32 tok/s | 1.22 s | $0.660 | $1.00 | 33k | – |
Price history
The price above, per 1M tokens, every day it changed.
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0006 | – | 11.6 s |
| Summarize a 30-page report | 12,000 / 600 | $0.0085 | – | 21.6 s |
| Code edit | 6,000 / 1,500 | $0.0055 | – | 51.6 s |
| Agentic coding session | 60,000 / 4,000 | $0.044 | – | 2.2 min |
| Structured extraction | 2,000 / 200 | $0.0015 | – | 8.3 s |
More from Alibaba
See also
Questions people ask
- What are Qwen2.5 Coder 32B Instruct's benchmark scores?
- As of 25 Oct 2026, Qwen2.5 Coder 32B Instruct scores 42.1 ±7.6 on the BenchLeader Index, strongest in reasoning (44) and weakest in coding (27). That rests on 7 independent benchmark results, and every score on this page links to its publisher.
- How does Qwen2.5 Coder 32B Instruct rank?
- Qwen2.5 Coder 32B Instruct is #609 of 785 ranked configurations on 25 Oct 2026. 386 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does Qwen2.5 Coder 32B Instruct cost per 1M tokens?
- As of 25 Oct 2026, Qwen2.5 Coder 32B Instruct costs $0.660 per 1M input tokens and $1.00 per 1M output tokens, $0.745 blended (three input tokens per output token). The list price last changed on 23 Oct 2026.
- When was Qwen2.5 Coder 32B Instruct released?
- Alibaba released Qwen2.5 Coder 32B Instruct on 18 Sept 2024.
Data as of 25 Oct 2026. Compare with another model.
Cite as: BenchLeader, “Qwen2.5 Coder 32B Instruct: benchmarks, pricing, speed and rank”, data as of 25 Oct 2026, https://www.benchleader.com/models/qwen2-5-coder-32b.