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 Aug 3, 2026Qwen2.5 Coder is a 32.8-billion-parameter coding specialist with a permissive Apache licence. Its leaderboard scores show clear strength in programming tasks alongside weaker general knowledge and creative writing results.
Pick this for budget coding inference where the output rate is identical across both tracked providers, or for Apache-licensed deployment where downloadable weights matter. Use it for instruction-following tasks that fit within a 32,768-token request limit. Skip it if you need strong creative writing, general reasoning measured at 13.2% on GPQA Diamond, or verified throughput beyond the single measured provider.
The case for it
- Coding specialty shows in relative leaderboard standing: its Arena Coding score sits 135.4 points above its own Creative Writing result.
- Fully permissive Apache 2.0 licence allows commercial use, fine-tuning and redistribution.
- Identical pricing across both tracked providers removes arbitrage complexity.
The case against it
- General knowledge and reasoning benchmarks are weak for its parameter scale: MMLU-Pro at 37.9% and GPQA Diamond at 13.2%.
- Creative writing is a clear relative weakness, with the lowest of its six Arena scores.
- Throughput is modest on the only measured provider at 19 tokens per second; the other provider is unverified in our data.
How good is it?
IntelligencePuzzles, maths, exam questions
Arena Text (overall)132nd of 143 · 1270.5
CodingWriting and fixing code on its own
Arena Coding124th of 143 · 1342.3
Arena Coding is the only board that has scored it for this.
AgenticPlanning, calling tools, staying on task
Qwen2.5 Coder 32B Instruct is not on Arena Agent (IPS), which is where the rating would come from, so there is no rating here. It is on SWE-bench Verified, in 39th of 39 with 9.
WritingWe do not rate this
Two boards come close and neither tests writing: Arena Creative Writing asks people which of two replies they prefer, and LiveBench Language tests whether a model understood a passage. So we show where Qwen2.5 Coder 32B Instruct placed and give it no mark out of five.
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, which is why they get no rating.
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?
- Fits in memory
- weights load entirely on the card
- Spills to system RAM
- some weights offload; much slower
- Too large
- will not load even with offload
- est
- size is calculated; the verdict could change by 10%
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.
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.
All 3 sizes here are calculated, not measured. We hold no measured file for this model, so each size comes from the parameter count and every verdict above inherits that uncertainty.
Check against your own machine → · All 71 devices, with every size →
Or rent it from someone else
Cheapest of 2 live listings. Picked at the widest standard context we hold, within one quantisation slice, so the numbers beside it are a price one host actually charges.
- per 1M tokens
- $0.66 in / $1.00 out
- Context served
- 33K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouter | $0.66 / $1.00 | 33K | not measured | Unknown | Unknown | Unknown |
| Cloudflare Workers AI | $0.66 / $1.00 | 33K | 13 tok/s | No | Yesunknown period | Unknown |
Cloudflare Workers AIAcross the 2 listings we hold: 1 say they do not train on prompts, 0 say they do and 1 do not say. 0 appear in the zero-retention registry we check; the rest are unknown to us rather than confirmed either way.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
- ✓
- Supported
- ✗
- Not supported
- Not published
- host gave no parameter list
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| OpenRouter | ✗ | ✗ | ✗ |
| Cloudflare Workers AI | ✗ | ✗ | ✗ |
Tool calling: 0 of 2 listings say yes, 2 say no. JSON output: 0 of 2 listings say yes, 2 say no. Strict schema: 0 of 2 listings say yes, 2 say no.
When we formed this view
Dates behind this page
Prices last checked 14h ago
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 do not say whether they train on prompts.
- We hold no cached-input rate for any of its listings.
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
- Modality record
- text->text
- Catalogue slug
- qwen-qwen2-5-coder-32b-instruct