Qwen2.5 Coder 32B Instruct
Qwen · released Nov 6, 2024 · Qwen/Qwen2.5-Coder-32B-Instruct
- Type
- Open weightsApache License 2.0
- Params
- 32.8B
- Context
- 33K
about 25K words of context
Our take
Written Sep 2, 2026Qwen2.5 Coder is a 32.8-billion-parameter coding specialist you can download and run yourself, released in late 2024 under a permissive Apache licence. It scores markedly better on coding chat than on general conversation, and its instruction-following accuracy is solid, though it struggles with graduate-level science and real-world software engineering tasks.
Pick this for coding assistance where the Arena Coding score is your guide, or for instruction-following workflows at 72.7% measured accuracy. Use it when you need Apache-licensed weights for commercial redistribution or self-hosting. Skip it if you need strong broad knowledge, graduate-level science reasoning, or end-to-end software engineering — its scores there are in single digits.
The case for it
- Coding chat score 72 points above its own general chat score, a genuine specialty.
- Apache 2.0 licence allows commercial use, fine-tuning and redistribution without restriction.
- Solid instruction-following accuracy at 72.7% on IFEval.
The case against it
- Weak on graduate-level science and broad knowledge: 13.2% on GPQA Diamond, 37.9% on MMLU-Pro.
- Only 9% of real-world GitHub issues resolved end-to-end on SWE-bench Verified.
- Throughput measured at 23 tokens per second on one host — the only figure we hold, with no basis to judge it against.
How good is it?
An open coding-focused model for developers, though it trails most models on everyday questions and prose.
- getting answers to everyday questionsArena Text (overall) · 156th of 168
- drafts, rewrites and editingArena Creative Writing · 161st of 168
- writing and completing codeArena Coding · 148th of 168
EverydayGeneral questions and everyday reasoning
Arena Text (overall)156th of 168 · 1271
CodingWriting and fixing code on its own
Arena Coding148th of 168 · 1342
Arena Coding is the only board that has scored it for this.
AgenticPlanning, calling tools, staying on task
Not yet scored on Arena Agent. It is on SWE-bench Verified, in 42nd of 42 with 9.
WritingDrafting and rewriting prose
Arena Creative Writing161st of 168 · 1207
Arena Creative Writing is the only board that has scored it for this.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done.
Every published score for this model10 scoresEvery figure we hold, from 10 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
GeForce RTX 4090 · 24 GB
Loads, but do not expect an assistant. Some of the weights sit in ordinary system memory, which is far slower than the card.20.7 GB of weights, plus 2.4 GB for the software that runs it and the smallest conversation it can hold, comes to 23.1 GB against the 22.8 GB this 24 GB device leaves free.
Comfortable fit
GeForce RTX 5090 · 32 GB
Room to spare. 7.8 GB spare means a 10% error in the size would not change the answer.
Apple M1 Pro (16-core GPU) · 32 GB
Borderline fit on an estimated size. It leaves 1 GB spare on a size we calculated rather than measured, and a 10% error either way would change the answer.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked 4 hours ago — each listing carries its own date.
- per 1M tokens
- $0.66 in / $1.00 out
- Context served
- 33K
- Throughput
- ~26 tok/s
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouterOpenRouter's own listing | $0.66 / $1.00checked 4 hours ago | 33K | not measured | Unknown | Unknown | Unknown |
| Cloudflare Workers AIThrough OpenRouter | $0.66 / $1.00checked 4 hours ago | 33K29K max reply | 26 tok/s | No | Yesunknown period | Unknown |
OpenRouterAcross the 2 listings we hold: 1 says it does not train on prompts, 0 say they do and 1 does not say. 0 appear in the zero-retention registry we check; the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| OpenRouterOpenRouter's own listing | ✗ | ✓ | ✗ |
| Cloudflare Workers AIThrough OpenRouter | ✗ | ✓ | ✗ |
Tool calling: 0 of 2 listings say yes, 2 say no. JSON output: 2 of 2 listings say yes. Strict schema: 0 of 2 listings say yes, 2 say no.
When we formed this view
Recent changes
What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 1 of 2 listings does not say whether it trains on prompts.
- We hold no cached-input rate for any of its listings.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
Licence and identifiers
What the licence allowsApache License 2.0, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Apache License 2.0
Fully permissive: commercial use, redistribution, and derivatives allowed. Requires attribution and a copy of the license. Includes an express patent grant.
Identifiers
- Hugging Face
- Qwen/Qwen2.5-Coder-32B-Instruct
- Architecture
- Dense
- Takes in, gives back
- Text in, text out
- Catalogue slug
- qwen-qwen2-5-coder-32b-instruct