DeepSeek V4.1 Flash OCR Benchmarked: 260 tok/s, Competitive Accuracy, Low Cost
solyarisoftware · x · 2026-09-10
NFTChen's test shows DeepSeek V4.1 Flash multimodal OCR achieves 6 errors + 1 miss on 270-character running script, tying GLM 5.3 Flash and Gemini 3.1 Pro; Kimi2.6 scores perfect, Qwen3.8-Max 5 errors. Speed: native multimodal stable at 260 tok/s (old version 100-150), single-image OCR under 3s with thinking off, others 15-50s. Cost-wise, Kimi/Qwen/GLM are accurate but expensive; DeepSeek V4.1 Flash after price cut balances accuracy and cost, ideal for large-scale ancient text OCR. Also, DeepSeek V4 Pro was replaced by V4.1 Flash after just 28 days, with Pro requests now routed to Flash pricing.
More from Models
- Kimi K3.1 Pro tops MathArena open models, closing frontier gap to about 6 months — teortaxesTex · 2026-10-26
- Ruff creator Charlie Marsh: frontier models find infinite bugs in any software — charliermarsh · 2026-10-26
- Dev testing one skill across Opus 5.5 and Sol 6.1 finds instruction following 'all over the map' — lucasmeijer · 2026-10-26
- supermemory unveils learner-1, claiming continual learning for any model — julianweisser · 2026-10-26
- Anthropic says Opus 5.5 cuts cost per task by ~40% vs Opus 5 — taufiqintech · 2026-10-26
- AI search APIs ranked: top 12 within 7 points, yet prices differ nearly 7x — FinanceYF5 · 2026-10-26