MaxxBench cost-quality chart: Muse Ember 1.0 hits 63.5% at $0.18, crushing rivals on value
AIatMeta · x · 2026-11-21
unLAP's MaxxBench score-vs-cost charts are making the rounds (retweeted by Meta's official account):
- Muse Ember 1.0: 63.5 score, $0.18/attempt, 211 tok/s — cheapest and fastest in the top score tier
- DeepSeek 1017: 59.0, $0.26, 147 tok/s
- Gemini 4 Argone: top score at 69.7, but $3.68 and only 29 tok/s
- Claude Haiku 5.5: 59.8 at $2.83; Qwen 4 Max: 57.9 at $1.50, clearly poor value
- Step 5 Preview: 42.7 at $2.96, behind on nearly every axis
Takeaway: on cost-adjusted coding performance, Muse Ember is currently a standout pick.
Related event: MaxxBench Charts Spark Debate as Muse Ember Tops Cost-Performance(2 posts)→
More from Models
- Downdetector reports OpenAI outage starting 3:13 PM EST as user complaints pile up — VoidStateKate · 2026-11-21
- Inkling 2 Mini on Ollama: MLX build, one-command hooks into Claude Code and OpenCode — ollama · 2026-11-21
- Inkling 2 Mini demoed building an RTX 5090 scene in Three.js — ollama · 2026-11-21
- Thinking Machines' Inkling-2-Mini trends on Hugging Face — thinkingmachines · 2026-11-21
- dottxt tool-call index: DeepSeek tops 100M daily calls, failure rates vary 0.04%–17% by provider — remilouf · 2026-11-21
- RL makes AI math proofs insanely hard to read as correctness rewards trump human style — burny_tech · 2026-11-21