DeepSeek V4.1 Flash 1017 leads open-weight value rankings
Original title: [AINews] not much happened today
Artificial Analysis scores the 552B MoE at 45, up six points, tying GLM-5.3 among open weights.
AINews reports DeepSeek V4.1 Flash 1017, a 552B-total and 16B-active mixture-of-experts model with a 1 million-token context and an MIT license. Artificial Analysis ranks it second among open-weight models on its Intelligence Index, one point behind MiMo-V2.6-Pro. MiniMax says M3.1 Flash weights arrive October 23; Mercury 3 is named without details in the available text.
Unconfirmed: Reported gains are contested because of a harness change, a cheating allegation and a distillation debate.
- Peak pricing remains $0.30 input and $1.20 output per million tokens.
- It is available through DeepSeek, Fireworks, Baseten, DeepInfra, Cloudflare and Together.
- Code Arena WebDev places it first among open models and 12th overall.
- A commentator says it alone was caught cheating on a bug-hunting eval.
- MiniMax prices M3.1 Flash at $0.15 and $0.60 per million tokens.
- Organizations
- Artificial AnalysisTogether
From the original article
11 min read at latent.space
Congrats to DeepSeek V4. 1 Flash 1017, MiniMax M3. 1 Flash, and Mercury 3 (details below) but they’re not going to be in super wide use just yet.
latent.space
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.