DeepSeek Ships V4 Pro, Rivaling Claude and GPT-5 for a Fraction of the Cost

DeepSeek Ships V4 Pro, Rivaling Claude and GPT-5 for a Fraction of the Cost Startup Fortune

DeepSeek's new AI model is by far the cheapest to run

DeepSeek's new AI model is by far the cheapest of well-known models to run, research firm says Indiatimes

DeepSeek Releases V4-Pro Model, Hikes Prices

DeepSeek Releases V4-Pro Model, Hikes Prices Silicon UK

😺 Google, OpenAI, DeepSeek dropped models today

😺 Google, OpenAI, DeepSeek dropped models today The Neuron

Deepseek ships improved V4 Pro, open-sources its agent software, and raises API prices

Deepseek has moved its flagship V4-Pro out of the testing phase and released its agent software, Harness v0.1, under the MIT license. API prices are going up at the same time, with cache hits jumping to six times their…

DeepSeek releases official V4 Pro model with sharply higher user prices

DeepSeek releases official V4 Pro model with sharply higher user prices Nikkei Asia

DeepSeek releases official V4 Pro model as it steps up expansion

DeepSeek releases official V4 Pro model as it steps up expansion Reuters

DeepSeek launches V4-Pro to rival Kimi K3 at lower prices

The Information: DeepSeek launches V4-Pro, its most advanced model that rivals Kimi K3 on some benchmarks at much lower prices, costing $0.44/1M input and $0.87/1M output tokens — Chinese AI developer DeepSeek…

DeepSeek plans substantial price increases

Bloomberg: DeepSeek says it plans to implement substantial price increases across its services; it currently charges $0.14/1M input and $0.28/1M output tokens for V4 Flash — DeepSeek plans to implement a…

DeepSeek V4-Flash costs $0.03 per test, vastly undercuts rivals

Eduardo Baptista / Reuters: Artificial Analysis: DeepSeek's V4-Flash costs $0.14/1M input and $0.28/1M output tokens, or $0.03 per test, far below Kimi K3's $0.86 and GPT-5.6 Sol's $1.86 — A version of Chinese…

DeepSeek's V4-Flash Undercuts OpenAI and Anthropic on Price Again

DeepSeek's V4-Flash Undercuts OpenAI and Anthropic on Price Again Startup Fortune

OpenAI says its new GPT 5.6 models are becoming more cost-efficient

OpenAI says its new GPT 5.6 models are becoming more cost-efficient BleepingComputer

New Deepseek Flash model matches OpenAI's GPT-5.6 Luna at roughly 60 percent lower cost

Deepseek's budget model V4 Flash gets a major boost with the "0731" update, jumping ten points to 50 on the Artificial Analysis Intelligence Index. That puts it just one point behind OpenAI's GPT-5.6 Luna, at roughly…

Enables Flash Attention for quantized V cache in DeepSeek V4 and MLA

llama : enforce the same K and V cache types for DeepSeek V4; enable FA if V cache is quantized (#25871) llama : enforce the same K and V cache types for DeepSeek V4; enable FA if V cache is quantized llama : enforce…

OpenAI slashes GPT-5.6 Luna prices by 80% as Chinese rivals rewrite the AI cost equation

OpenAI slashes GPT-5.6 Luna prices by 80% as Chinese rivals rewrite the AI cost equation Startup Fortune

DeepSeek V4 Flash now runs updated weights on AI Gateway

now runs on updated weights by default on AI Gateway, with notably stronger agentic capabilities. On Terminal-Bench, it scores 82.7, up 25.8 points from 56.9 in the April preview.DeepSeek V4 Flash Requests to pick up…

Open-weight models match four-month-old frontier cyber performance

The British AI Security Institute warns that open-weight models like GLM-5.2 and DeepSeek V4-Pro now trail closed frontier models in cyber capabilities by four to seven months. At the start of 2025, the gap was still…

Moonshot’s Kimi K3 pushes Chinese AI into Fable-level territory

Moonshot’s Kimi K3 pushes Chinese AI into Fable-level territory Fortune

DeepSeek’s 75% price cut pressures AI market, impacts Anthropic valuation

DeepSeek’s 75% price cut pressures AI market, impacts Anthropic valuation Crypto Briefing

Exclusive | Anthropic, China and why Pax Silica architect thinks the US can keep AI lead

Exclusive | Anthropic, China and why Pax Silica architect thinks the US can keep AI lead South China Morning Post

Adds Gemma 4 12B Unified, Sapiens2, and DeepSeek-OCR-2 models

Release v5.10.1 v5.10.0 was yanked as we publish on a corrupted branch. Sorry everyone, this happens when we rush a release!!! New Model additions Gemma4 unified+ Gemma4 MTP Gemma 4 12B Unified is an encoder-free…

Adds NVFP4 MoE, CUDA graphs, and MTP for DeepSeek V4

Highlights This release features 459 commits from 230 contributors (63 new)! DeepSeek V4 maturity: DeepSeek V4 received a major hardening pass this cycle — the model was reorganized into a dedicated…

Deepseek wants to take on Claude Code and OpenAI's Codex with "Deepseek Code"

Deepseek is building a new team in Beijing to develop its own AI code agent, working title "Deepseek Code," a direct competitor to Claude Code, Codex, and Cursor. Applicants should know agent loops, MCP, and context…

Open model bonanza: Gemma 4, DeepSeek V4, Kimi K2.6 debut

An eventful month with one flagship release after another

DeepSeek V4: The Open-Source Model Frontier Labs Feared

Transformers v5.8.0

Release v5.8.0 New Model additions DeepSeek-V4 DeepSeek-V4 is the next-generation MoE (Mixture of Experts) language model from DeepSeek that introduces several architectural innovations over DeepSeek-V3. The…

vLLM v0.20.0

vLLM v0.20.0 Highlights This release features 752 commits from 320 contributors (123 new)! DeepSeek V4: Initial DeepSeek V4 support landed (#40860), with DSML token-leakage fix in DSV4/3.2 (#40806), DSA + MTP IMA fix…

vLLM v0.11.0

Highlights This release features 538 commits, 207 contributors (65 new contributors)! This release completes the removal of V0 engine. V0 engine code including AsyncLLMEngine, LLMEngine, MQLLMEngine, all attention…

DeepSeek V4 Pro 0813 vs GPT-5.6 Sol on DeepSWE: Cost, Coding, and Routing

We ran 904 DeepSWE rollouts on DeepSeek V4 Pro 0813 and GPT-5.6 Sol. Sol leads pass@1 by 10 points at 35x the cost; Pro wins pass@4, and a Pro-first cascade hits 83.0%.

DeepSeek V4 Pro 0813 vs Claude Fable 5 on DeepSWE: Cost, Coding, and Routing

We ran 904 DeepSWE rollouts on DeepSeek V4 Pro 0813 and Claude Fable 5. Fable leads pass@1 at 90x the cost; Pro wins pass@4, and a Pro-first cascade hits 82.7%.