| Model | Score | Status | Released | Price in/out | Context | |
|---|---|---|---|---|---|---|
| DeepSeek V4.1-Flash agentic coding and vision agents at Flash-tier cost | 8.6 | GA | 2026-09-10 | $0.30 / $1.20 | 1M | Review → |
| DeepSeek V4-Pro frontier-grade agentic coding at open-weights cost | 8.5 | GA | 2026-08-13 | $1.32 / $3.96 | 1M | Review → |
| DeepSeek V4-Flash frontier-adjacent quality at the lowest cost in market | 8.7 | retired | 2026-04-24 | $0.14 / $0.28 | 1M | Review → |
| DeepSeek V3.2 open-weights math/reasoning at GA stability | 8.2 | GA | 2025-12-01 | $0.25 / $0.38 | 128K | Review → |
| DeepSeek V3.1 the model that put DeepSeek on the production-agent map | 7.7 | GA | 2025-08-21 | $0.21 / $0.79 | 128K | Review → |
| DeepSeek R1 (0528) exposed-CoT reasoning at a fraction of o-series cost | 8.0 | GA | 2025-05-28 | $0.55 / $2.19 | 128K | Review → |
| DeepSeek-VL2 open-weights OCR and document understanding at the edge | 7.2 | GA | 2024-12-13 | — / — | 4K | Review → |
DeepSeek retired V4-Flash on its API on 2026-09-10 in favour of its successor, DeepSeek V4.1-Flash, and the deepseek-v4-flash model name now routes to V4.1-Flash at V4.1-Flash rates ($0.30 in / $1.20 out per 1M tokens at peak, half that off-peak).
DeepSeek V4-Flash is the volume tier of the V4 family — a 284B-parameter Mixture-of-Experts model that activates just 13B parameters per token, inherits V4-Pro's 1M-token context and CSA/HCA sparse attention; DeepSeek listed it at $0.14 in / $0.28 out per 1M tokens from its preview until mid-August 2026. It shipped as a preview on 2026-04-24 with MIT-licensed open weights, got its official release (DeepSeek-V4-Flash-0731) on 2026-07-31, and was the default model behind the legacy deepseek-chat (non-thinking) and deepseek-reasoner (thinking) aliases until they retired on 2026-07-24. The single sentence a buyer needs: for high-throughput chat, agent inner loops, and batch processing where per-token cost is the binding constraint, V4-Flash was the cheapest frontier-adjacent model in the market by a wide margin, and its open weights keep it available after the API retirement.