DeepSeek V4.1 Flash
The #1 open-weight model (Vals AI) — 552B MoE, 1M context, MIT.
About
DeepSeek V4.1 Flash is DeepSeek's lightweight flagship: a 552B mixture-of-experts model (8B active prefill / 16B decode) with native multimodal input, 1M-token context, and MIT-licensed weights.
Released September 10, 2026; independent evaluator Vals AI ranks it the #1 open-weight model. API: $0.30 in / $1.20 out per 1M at peak. The open-weights answer to the frontier — matches or beats V4-Pro on agentic coding at a fraction of the serving cost.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Vals AI: #1 open-weight model. Artificial Analysis: Intelligence Index 50 (V4-Flash-0731). Provider claims 82.7 Terminal Bench 2.1 — provider-claimed.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 9/10
Scored 9/10 (verified tier). MIT open weights, 1M context, independently ranked #1 open-weight model (Vals AI), live API. One of the highest-trust open models in the catalog. Docked only for provider-claimed benchmark figures. Verified tier.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 9.0/10
- Maintenance · recency, releases, issue load
- 9.0/10
- Openness · license + docs
- 9.4/10
- Manual review · sanity read + red-flag check
- 9.0/10
- · provider-tier estimate — thinner evidence (no public adoption metric)
- · Vals AI #1 open-weight model; MIT; 1M context
- · released 33 days ago
- · manual review: V4.1-Flash — 552B MoE, 8B/16B active, native multimodal encoder-decoder; MIT weights; $0.30/$1.20 peak per 1M; license: MIT