QwQ-32B APIs – o1 like reasoning at 1% the cost
Ubicloud is an open source alternative to AWS. Today, we launched our inference APIs, built with open source AI models. QwQ-32B-Preview is one of those models; and it can provide o1-like reasoning at 1% the cost. QwQ is licensed under Apache 2.0 [1] and Ubicloud under AGPL v3. We deploy open models on a cloud stack that can run anywhere. This allows us to offer great price / performance. From an accuracy standpoint, QwQ does well in math and coding domains. For example, in the MMLU-Pro Computer Science LLM Benchmark, the accuracy rankings are as follows. Claude-3.5 Sonnet (82.5),…
In plain words
Ubicloud offers inference APIs powered by open source AI models, providing an alternative to commercial cloud services. QwQ-32B-Preview delivers advanced reasoning capabilities comparable to o1 at significantly reduced costs, with particular strength in mathematics and computer science tasks. The service runs on open source infrastructure deployable anywhere, offering competitive pricing at $0.60 per million tokens. QwQ is Apache 2.0 licensed, making it suitable for developers and organizations seeking cost-effective AI inference with strong performance on technical problems.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
Ubicloud is an open source alternative to AWS. Today, we launched our inference APIs, built with open source AI models. QwQ-32B-Preview is one of those models; and it can provide o1-like reasoning at 1% the cost. QwQ is licensed under Apache 2.0 [1] and Ubicloud under AGPL v3. We deploy open models on a cloud stack that can run anywhere. This allows us to offer great price / performance. From an accuracy standpoint, QwQ does well in math and coding domains. For example, in the MMLU-Pro Computer Science LLM Benchmark, the accuracy rankings are as follows. Claude-3.5 Sonnet (82.5), QwQ-32B-Preview (79.1), and GPT 4o 2024-11-20 (73.1). [2] You can start evaluating QwQ (and Llama 3B / 70B) by logging into the Ubicloud console: https://console.ubicloud.com/create-account We also provide an AI chat box for convenience. We price the API endpoints at $0.60 per M tokens, or 100x lower than o1’s output token price. Also, when using open models, your first million tokens each month are free. This way, you can start evaluating these models today. ## OpenAI o1 or QwQ-32B In math and coding benchmarks, QwQ-32B ties with o1 and outperforms Claude 3.5 Sonnet. In our qualitative tests, we found o1 to perform better. For example, we asked both models to “add a pair of parentheses to the incorrect equation: 1 + 2 * 3 + 4 * 5 + 6 * 7 + 8 * 9 = 479, to make the equation true.” [3] QwQ’s answer shows iterative reasoning steps, where the model enumerates over answers using light heuristics. o1’s answer to the same question feels like an iterative deepen-and-test (though not purely depth-first). When we asked the models harder questions, it felt that o1 could understand the question better and employ more complex strategies. [3][4] Finally, we found that o1’s advantage in reasoning compounded with other ones. For example, we asked both models to write example Python programs. Looking at the answers, it became clear that o1 was trained on a larger data set and that it was aware of Python libraries that QwQ-32B didn’t know about. Further, QwQ-32B at times flip flopped between English and Chinese, making it harder for us to understand the model. [3] Now, if we think that o1 has these advantages, why the heck are we doing a Show HN on QwQ-32B (and other open weight models)? Two reasons. First, QwQ is still comparable to o1 and Ubicloud offers it for 100x less. You can employ a dozen QwQ-32Bs, prompt them with different search strategies, use VMs to verify their results, and still come in under what o1 costs. In the short term, combining these classic AI search strategies with AI models feels much more efficient than trying to “teach” an uber AI model. Second, we think open source fosters collaboration and trust -- and that is its superpower that compounds over time. We foresee a future where open source AI not only delivers top-quality results, but also surpasses proprietary models in some areas. If you believe in that future and are looking for someone to partner with on the infrastructure side, please hit us up at [email protected]! [1] https://qwenlm.github.io/blog/qwq-32b-preview/ [2] https://huggingface.co/blog/wolfram/llm-comparison-test-2024... [3] https://github.com/ubicloud/ubicloud/discussions/2608 [4] https://arxiv.org/pdf/2201.11903
Does the same job
all alternatives →
- Q2Qwen-2.5-32B is now the best open source OCR model2025 · github.com · ▲211
Last week was big for open source LLMs. We got: - Qwen 2.5 VL (72b and 32b) - Gemma-3 (27b) - DeepSeek-v3-0324 And a couple weeks ago we got the new mistral-ocr model. We updated our OCR benchmark to include the new models. We evaluated 1,000 documents for JSON extraction accuracy. Major takeaways: - Qwen 2.5 VL (72b and 32b) are by far the most impressive. Both landed right around 75% accuracy (equivalent to GPT-4o’s performance). Qwen 72b was only 0.4% above 32b. Within the margin of error. - Both Qwen models passed mistral-ocr (72.2%), which is specifically trained for OCR. - Gemma-3…
General ComputeMay 2026 · generalcompute.com · ▲315AI models that run on an inference cloud optimized for speed
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, January 2025
the whole month →- IM
Hello! I'm Byran. I spent the past ~6 months engineering a laptop from scratch. It's fully open-source on GH at: https://github.com/Hello9999901/laptop
Dev tools · 2025 · byran.ee
- TITetris in a PDF▲1,289
I realized that the PDF engines of modern desktop browsers (PDFium and PDF.js) support JavaScript with enough I/O primitives to make a basic game like Tetris. It was a bit tricky to find a union of features that work in both engines, but in the end it turns out that showing/hiding annotation "fields" works well to make monochrome pixels, and keyboard input can be achieved by typing in a text input box. All in all it's quite janky but a nice reminder of how general purpose PDF scripting can be. The linked PDF is all ASCII so you can just open it in a text editor, or have a look at…
Life & fun · 2025 · th0mas.nl
Create lifelike, personalized AI avatars from text prompts
AI · 2025 · jogg.ai
