Qwen3.8 Beats Fable in Coding Benchmarks
0xsachi · x · 2026-08-03
Developer @0xsachi tweeted that Alibaba's open-weights Qwen3.8 model has outperformed Fable in coding benchmarks, celebrating the continued advantage of open-source models.
More from Models
- Benchmark Manipulation: You Can Reach Any Conclusion by Controlling Tests — felix_red_panda · 2026-08-03
- Teknium: DeepSeek V4 Flash Hits 70 tok/s on Local Inference — Teknium · 2026-08-03
- Thinking Machines Releases 276B MoE Model; Simile Raises $200M — thione · 2026-08-03
- Thinking Machines Releases Open-Weight Inkling-Small: 276B MoE, 12B Active, 1M Context — thione · 2026-08-03
- Anthropic Releases Claude Opus 5: New SOTA for Coding and Knowledge Work at Half Price — thione · 2026-08-03
- Fish Audio Raises $52M Seed, Launches S2.1 Pro Voice Model — thione · 2026-08-03