In 30 seconds
In an NVIDIA-partnered 2025 short, Matthew Berman describes gaming on an RTX 5090 and tests an 8-billion-parameter DeepSeek R1 distilled model in LM Studio. He reports 78 tokens per second for one run; the short does not establish performance across models or workloads. Berman also discloses a small investment in LM Studio.
- 1Partnered hardwareBerman says NVIDIA provided the RTX 5090, invited him to CES and partnered on the video.
- 2Gaming firstHe says he tried Cyberpunk at maximum settings and enjoyed how Path of Exile 2 looked.
- 3Local AI generationAn 8B distilled DeepSeek R1 model generates Python Snake code in LM Studio, with a reported 78 tokens per second on that run.
Takeaways
- NVIDIA supplied the RTX 5090 and partnered with Berman on the video.
- Berman runs an 8B DeepSeek R1 distilled model locally in LM Studio.
- He asks for a Python Snake game and reports 78 tokens per second for that run.
- The short does not provide a controlled benchmark across hardware, models or configurations.
Read the edited text by chapter
Games and a local model test
Berman introduces the RTX 5090 and his NVIDIA partnership. He reports impressions from games and the speed of a single local AI generation.
NVIDIA sent me an RTX 5090, its high-end next-generation GPU, and invited me to CES, where Jensen Huang introduced it. NVIDIA also partnered with me on this video. My first test was gaming: I say Cyberpunk ran at maximum settings without a problem, and Path of Exile 2 looked gorgeous.
For AI, I load the 8-billion-parameter DeepSeek R1 distilled model locally in LM Studio and ask it to write Snake in Python. It shows its reasoning and produces code; I report 78 tokens per second for this run. This single example does not establish a general hardware ranking or show whether the game was tested. I disclose a small investment in LM Studio in the video description.
More from this channel
34:21EN
Matthew Berman· Oct 27, 2026Matthew Berman Examines Anthropic’s IPO Filing, AI Growth and GovernanceMatthew Berman, Sam Altman · Forward FutureThe presenter reviews the S-1’s financial metrics, compute strategy and voting structure while questioning the tension between rapid AI development and safety.
0:36ENMatthew Berman· Oct 25, 2026Kimi K3.1 and a Rubik’s Cube Solver DemonstrationMatthew Berman · Forward FutureMatthew Berman shows a scrambled virtual cube being solved and gives the Kimi K3.1 test a favorable assessment.
25:50ENMatthew Berman· Oct 24, 2026Kimi K3.1 Pro Tested: Strong 3D Demos, Uneven Results and Competitive PricingMatthew Berman · Forward FutureMatthew Berman tests the open-weight model on coding, 3D rendering and visual explanations, then compares benchmark results and prices.
32:20ENMatthew Berman· Oct 23, 2026Who Is Local AI For? The Costs, Trade-offs and a Hybrid FutureMatthew Berman · Forward FutureMatthew Berman weighs data control and independence against hardware cost, setup complexity and inference speed.
1:03:37ENMatthew Berman· Oct 23, 2026Matthew Berman’s Team Tests Kimi K3.1 on Animation, Games, and ChartsMatthew Berman, Alex, Brian · Forward FutureA live team AMA turns into an informal test of Kimi K3.1, from personal-agent workflows to motion graphics, a browser game, and a benchmark chart.
0:53ENMatthew Berman· Oct 22, 2026Sam Altman on AI and the Discovery of New KnowledgeSam Altman · OpenAIAltman expects progress in mathematical discovery to extend to biology, materials and energy.