LongCat-2.5 vs Qwen3.8-Max: same browser game built for $0.28 vs $0.82
omarsar0 · x · 2026-11-03
omarsar0 tested LongCat-2.5 (Meituan) against Qwen3.8-Max with the same one-line prompt: build a playable Angry Birds-style browser game. Both worked; LongCat's version had 3 levels, 3 bird types, breakable wood/stone/glass blocks, sound effects and star ratings.
- LongCat-2.5: 34 minutes, $0.28
- Qwen3.8-Max: $0.82 — nearly 3x pricier for the same task
- LongCat-2.5: 1.6T total params, 48B active per token, native multimodal input, trained on sequences up to 1M tokens, aimed at long-horizon tasks
His takeaway: a strong cheap option for full app builds.
More from coding & agent
- oxc parser gets up to 20% faster in next release, cutting rolldown bundling time ~6% — irvinebroque · 2026-11-03
- Qwen's Astra gains come from an agentic ML loop, not weight introspection — petrusenko_max · 2026-11-03
- DeepSeek's super-agent is 1.9x faster but burns 3.8x the tokens — AIFlow_ML · 2026-11-03
- TMax v1.1 released: open terminal agents under 30B gain up to 6 points on TB2 — natolambert · 2026-11-03
- Voice Ring: open-source voice-reactive light ring with sound direction and emotion recognition — alexcovo_eth · 2026-11-03
- Skipping YAML: encoding config values into a Qwen 0.8B model instead — chrisalbon · 2026-11-03