Qwen 2.5 Coder/Max is currently the top open-source model for coding, with the highest HumanEval (~70–72%), LiveCodeBench (70.7), and Elo (2056) scores among open models.
DeepSeek V3/Coder V2 remains strong, especially in reasoning/math, but is slightly behind Qwen in code generation and competitive programming Elo.
Meta Llama 4 Maverick offers unmatched context length (up to 10 million tokens) and robust general coding, but its coding benchmarks are a bit lower than Qwen and DeepSeek.
Open Source LLM Progress & Leadership
Open Source Progress: The open-source LLM ecosystem is advancing rapidly, with Qwen, DeepSeek, and Meta pushing the boundaries in both code and general language tasks. Coding benchmarks show open models now rival or surpass many closed models from just a year ago.
Current Leader: Qwen 2.5 Coder/Max is considered the open-source leader for coding as of May 2025, based on benchmark dominance and real-world developer feedback
DeepSeek V3 vs. Qwen3: Coding Task Performance
Benchmark Results and User Feedback
Qwen3 (especially the flagship Qwen3-235B-A22B model) consistently outperforms DeepSeek V3 (and DeepSeek R1) in coding tasks across multiple benchmarks and real-world evaluations
* On LiveCodeBench (code generation), Qwen3-235B scored 70.7, which is higher than DeepSeek R1 and most other open-source models except for some proprietary models like Gemini
* CodeForces Elo, a competitive programming benchmark, shows Qwen3-235B at 2056—again, ahead of DeepSeek R1 and other major competitors
* On HumanEval (coding), Qwen3 models have been reported to beat DeepSeek R1 and even some OpenAI models in code generation quality and correctness
User Experience
Developers report that Qwen3 produces more functional, user-friendly, and well-structured code than DeepSeek V3, often replacing even top-tier proprietary models for coding workflows
* Qwen3 is praised for its efficiency, with fewer active parameters needed for strong results, making it more accessible for local deployment
DeepSeek V3 Strengths
While Qwen3 leads in most coding benchmarks, DeepSeek V3 may have a slight edge in certain complex multi-step mathematical reasoning tasks, but this advantage does not generally extend to coding
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
