DeepSeek open-sources inference optimizations with 60–85% faster generation [pdf]

Re:DeepSeek open-sources inference optimizations with 60–85%... by piterrro on Saturday June 27, 2026 @06:09AM (#48696874)

I’ve been using DeepSeek v4 pro for a month now in Kilo Code and its great. Fast, reliable, large context window and cheap as… Did 1,5B tokens this month and cost me 40usd (majority cached, but still).

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by spiderfarmer on Saturday June 27, 2026 @06:16AM (#48696901)

Is there a way to see how many tokes one does with claude code (pro)?

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by bpavuk on Saturday June 27, 2026 @06:36AM (#48697011)

the casino has no clocks, as one HN user put it some time ago.

I second ccusage, it's nice

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by cptchaos on Saturday June 27, 2026 @06:20AM (#48696927)

https://ccusage.com/

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by edg5000 on Saturday June 27, 2026 @07:09AM (#48697208)

It's in the JSONs in ~/.claude, but last 30 days only I think. You can have the model analyze history. So for correct history you'd need to run history analysis on a cron job or something. Kinda hacky.

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by Stagnant on Saturday June 27, 2026 @12:25PM (#48699570)

The 30 day limit can be overridden by adding "cleanupPeriodDays": 9999 to .claude/settings.json

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by O_H_E on Saturday June 27, 2026 @08:38PM (#48703207)

https://github.com/kenn-io/agentsview

> Local-first session search, analytics, insights, and token use statistics for coding agents, supporting Claude Code, Codex, and more than 20 other agents.

solid piece of software

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by richardlblair on Saturday June 27, 2026 @08:26AM (#48697652)

I've been using omp with deepseek as my task and quicktask agents, and sonnet as everything else.

It's drastically reduced my AI spend. I went from spending $40/day to $10/day.

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by throwa356262 on Saturday June 27, 2026 @03:38PM (#48701048)

Have you tried reasonix?

https://github.com/esengine/deepseek-reasonix

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by fer on Saturday June 27, 2026 @10:36AM (#48698658)

Which provider? I went through 40 bucks on it on openrouter. It was not a lot of back and forth, context ended at around 300k, 15kloc output. I was using opencode, unsure if I can make the total token count visible.

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by peheje on Saturday June 27, 2026 @11:24AM (#48699054)

OpenRouter sometimes chooses a very expensive provider. Try the floor slug or choose directly the provider. I moved to just putting 5 dollars directly on deepseek instead of going through OR.

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by apitman on Saturday June 27, 2026 @10:45AM (#48698754)

Have you compared Kilo to Pi or OpenCode? Those are the two I'm most familiar with but always looking for alternatives.

[ Parent | on HN ]

Re:DeepSeek open-sources inference optimizations with 60–85%... by redman25 on Monday June 29, 2026 @04:22PM (#48724645)

I've been preferring Mimo recently. Same price as deekseek, more reliable tool calling (subjectively), and has some nice qualities in terms of prose, etc.

I've heard others say that Deepseek tends to be smarter on specific problems but that Mimo tends to more well-rounded.

[ Parent | on HN ]