{"slug": "reasonix", "name": "Reasonix", "category": "ai-coding", "category_name": "AI coding tools", "description": "Coding agent harness for DeepSeek models, built to maximize prefix-cache hits.", "recommenders": 5, "commenters": 6, "warners": 0, "score": 5.021, "rising": 0.36, "first": "2026-05-24", "latest": "2026-09-23", "why": "Recommended as a DeepSeek agent harness whose prefix-cache alignment yields mostly cache-hit tokens, letting commenters run long coding sessions and large code analyses for pennies.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 7, "neutral": 2, "compare": 0, "warn": 0}, "spark": [0, 0, 0, 0, 0, 0, 0, 0, 5, 1, 1, 0], "listed": true, "loved": ["Low cost", "Cache hits"], "caveat": "Data to China", "short_quote": "When I want a harness to inhale a codebase and give me an answer, Reasonix has been the go-to for a while now, and it's great.", "icon": false, "domain": null, "github_repo": null, "official_url": "", "pricing": "unknown", "aliases": ["Reasonix", "reasonix"], "verdict": [{"text": "Recommended as a DeepSeek agent harness whose prefix-cache alignment yields mostly cache-hit tokens,", "cites": [{"id": 797043538, "item_id": 49815221, "url": "https://news.ycombinator.com/item?id=49815221", "by": "abalashov", "date": "2026-09-23", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly", "excerpt": "Data points from my own usage, for whatever they're worth[1]. I'm just an individual contributor and hardly represent enterprise or large orgs: I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! 1) For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit i", "context": "individual contributor", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "The current balance of power in open models"}, {"id": 796953682, "item_id": 49809605, "url": "https://news.ycombinator.com/item?id=49809605", "by": "abalashov", "date": "2026-09-22", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache.", "excerpt": "Welcome to the dark side. I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit input tokens, so very cheap large-scale code analyses and things that require mega context windows (at the cost of some at", "context": "", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis"}, {"id": 785488995, "item_id": 49093062, "url": "https://news.ycombinator.com/item?id=49093062", "by": "wilj", "date": "2026-07-28", "stance": "recommend", "quote": "And Reasonix + Deepseek is mostly cache hits when I let it grind on it for a long time", "excerpt": "Similar here. I use Claude Code and Codex for developing specs, but then have Deepseek v4 Pro implement because it just seems to actually do what it's told, given enough tries. And Reasonix + Deepseek is mostly cache hits when I let it grind on it for a long time, and I can have a huge fleet of agents with very specific focus for code review and rule enforcement mostly sharing cache.", "context": "", "use_case": "agent fleet for code review", "switched_from": "", "story": "Boris Cherny says \"delete your Claude.md every 6 months\""}, {"id": 783711504, "item_id": 48981969, "url": "https://news.ycombinator.com/item?id=48981969", "by": "arikrahman", "date": "2026-07-20", "stance": "recommend", "quote": "I use the reasonix or whale harness for Deepseek, prefer the hyper optimized cache hits making request converge on almost free.", "excerpt": "This is why I use the reasonix or whale harness for Deepseek, prefer the hyper optimized cache hits making request converge on almost free.", "context": "", "use_case": "DeepSeek harness", "switched_from": "", "story": "Stop Using OpenCode"}, {"id": 780316016, "item_id": 48769751, "url": "https://news.ycombinator.com/item?id=48769751", "by": "arikrahman", "date": "2026-07-02", "stance": "recommend", "quote": "I use Reasonix with planning-with-files and that's been better than even openspec.", "excerpt": "Likewise. I use Reasonix with planning-with-files and that's been better than even openspec. Cache hits with the help of Reasonix on Deepseek make requests practically free. And that's unsubsidized, with American providers like Digital Ocean or cloudflare.", "context": "", "use_case": "AI coding with DeepSeek", "switched_from": "", "story": "Superpowers 6"}]}, {"text": " letting commenters run long coding sessions and large code analyses for pennies.", "cites": [{"id": 785986657, "item_id": 49124166, "url": "https://news.ycombinator.com/item?id=49124166", "by": "bwfan123", "date": "2026-07-31", "stance": "recommend", "quote": "With reasonix or pi, I can code all day long and pay a few pennies for it.", "excerpt": "> For the Code Agent tasks among the public benchmarks above, DeepSeek-V4-Flash-0731 is evaluated with the minimal mode of DeepSeek Harness (to be released) as the agent framework So, are they planning to announce an optimized coding agent harness as well ? DSv4 flash is a fantastic model, and my daily driver. With reasonix or pi, I can code all day long and pay a few pennies for it. No token anxiety. Whereas the sam", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis"}, {"id": 783398945, "item_id": 48962434, "url": "https://news.ycombinator.com/item?id=48962434", "by": "SwellJoe", "date": "2026-07-18", "stance": "recommend", "quote": "with Reasonix, DeepSeek is free real estate. Seems to just go and go for pennies.", "excerpt": "Yep, with Reasonix, DeepSeek is free real estate. Seems to just go and go for pennies. And, DeepSeek is what I use for any task that works best with an API. It's cheap enough to where I don't think about cost, made even cheaper by DeepSeek having the most effective and cheap caching in the industry, and it's good enough to where I rarely have to follow up with a more expensive model or manually fix things. It's been ", "context": "", "use_case": "running DeepSeek cheaply", "switched_from": "", "story": "The Kimi K3 Moment"}, {"id": 797043538, "item_id": 49815221, "url": "https://news.ycombinator.com/item?id=49815221", "by": "abalashov", "date": "2026-09-23", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly", "excerpt": "Data points from my own usage, for whatever they're worth[1]. I'm just an individual contributor and hardly represent enterprise or large orgs: I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! 1) For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit i", "context": "individual contributor", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "The current balance of power in open models"}, {"id": 796953682, "item_id": 49809605, "url": "https://news.ycombinator.com/item?id=49809605", "by": "abalashov", "date": "2026-09-22", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache.", "excerpt": "Welcome to the dark side. I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit input tokens, so very cheap large-scale code analyses and things that require mega context windows (at the cost of some at", "context": "", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis"}]}, {"text": " Few complaints appear; the main caveat is that it requires sending data to China,", "cites": [{"id": 796953682, "item_id": 49809605, "url": "https://news.ycombinator.com/item?id=49809605", "by": "abalashov", "date": "2026-09-22", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache.", "excerpt": "Welcome to the dark side. I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit input tokens, so very cheap large-scale code analyses and things that require mega context windows (at the cost of some at", "context": "", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis"}]}, {"text": " with some attentional drift in very large contexts and open questions about cache benefits via OpenRouter.", "cites": [{"id": 796953682, "item_id": 49809605, "url": "https://news.ycombinator.com/item?id=49809605", "by": "abalashov", "date": "2026-09-22", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache.", "excerpt": "Welcome to the dark side. I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit input tokens, so very cheap large-scale code analyses and things that require mega context windows (at the cost of some at", "context": "", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis"}, {"id": 797043538, "item_id": 49815221, "url": "https://news.ycombinator.com/item?id=49815221", "by": "abalashov", "date": "2026-09-23", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly", "excerpt": "Data points from my own usage, for whatever they're worth[1]. I'm just an individual contributor and hardly represent enterprise or large orgs: I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! 1) For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit i", "context": "individual contributor", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "The current balance of power in open models"}, {"id": 772126704, "item_id": 48257919, "url": "https://news.ycombinator.com/item?id=48257919", "by": "schaefer", "date": "2026-05-24", "stance": "neutral", "quote": "If I use Reasonix to talk to a deepseek endpoint through openrouter", "excerpt": "Okay, I'm curious. From the FAQ, I see: >Can I point it at a self-hosted / private DeepSeek endpoint? >Yes. Since 0.30 we accept non-standard key prefixes for self-hosted DeepSeek endpoints. Just point `baseUrl` at your internal address \u2014 the loop, cache strategy, and tool protocol are unchanged. But my question is: If I use Reasonix to talk to a deepseek endpoint through openrouter, am I still getting the cache-hit ", "context": "", "use_case": "DeepSeek agent harness", "switched_from": "", "story": "DeepSeek reasonix, engineered for low token cost during long running sessions"}]}], "praise": [{"label": "Low cost", "summary": "Commenters say DeepSeek through Reasonix costs pennies, making requests nearly free.", "quotes": [{"id": 785986657, "item_id": 49124166, "url": "https://news.ycombinator.com/item?id=49124166", "by": "bwfan123", "date": "2026-07-31", "stance": "recommend", "quote": "With reasonix or pi, I can code all day long and pay a few pennies for it.", "excerpt": "> For the Code Agent tasks among the public benchmarks above, DeepSeek-V4-Flash-0731 is evaluated with the minimal mode of DeepSeek Harness (to be released) as the agent framework So, are they planning to announce an optimized coding agent harness as well ? DSv4 flash is a fantastic model, and my daily driver. With reasonix or pi, I can code all day long and pay a few pennies for it. No token anxiety. Whereas the sam", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis"}, {"id": 783398945, "item_id": 48962434, "url": "https://news.ycombinator.com/item?id=48962434", "by": "SwellJoe", "date": "2026-07-18", "stance": "recommend", "quote": "with Reasonix, DeepSeek is free real estate. Seems to just go and go for pennies.", "excerpt": "Yep, with Reasonix, DeepSeek is free real estate. Seems to just go and go for pennies. And, DeepSeek is what I use for any task that works best with an API. It's cheap enough to where I don't think about cost, made even cheaper by DeepSeek having the most effective and cheap caching in the industry, and it's good enough to where I rarely have to follow up with a more expensive model or manually fix things. It's been ", "context": "", "use_case": "running DeepSeek cheaply", "switched_from": "", "story": "The Kimi K3 Moment"}, {"id": 780316016, "item_id": 48769751, "url": "https://news.ycombinator.com/item?id=48769751", "by": "arikrahman", "date": "2026-07-02", "stance": "recommend", "quote": "I use Reasonix with planning-with-files and that's been better than even openspec.", "excerpt": "Likewise. I use Reasonix with planning-with-files and that's been better than even openspec. Cache hits with the help of Reasonix on Deepseek make requests practically free. And that's unsubsidized, with American providers like Digital Ocean or cloudflare.", "context": "", "use_case": "AI coding with DeepSeek", "switched_from": "", "story": "Superpowers 6"}]}, {"label": "Cache hits", "summary": "Its alignment to DeepSeek's prefix cache yields 95%+ cache-hit input tokens on long runs.", "quotes": [{"id": 796953682, "item_id": 49809605, "url": "https://news.ycombinator.com/item?id=49809605", "by": "abalashov", "date": "2026-09-22", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache.", "excerpt": "Welcome to the dark side. I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit input tokens, so very cheap large-scale code analyses and things that require mega context windows (at the cost of some at", "context": "", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis"}, {"id": 785488995, "item_id": 49093062, "url": "https://news.ycombinator.com/item?id=49093062", "by": "wilj", "date": "2026-07-28", "stance": "recommend", "quote": "And Reasonix + Deepseek is mostly cache hits when I let it grind on it for a long time", "excerpt": "Similar here. I use Claude Code and Codex for developing specs, but then have Deepseek v4 Pro implement because it just seems to actually do what it's told, given enough tries. And Reasonix + Deepseek is mostly cache hits when I let it grind on it for a long time, and I can have a huge fleet of agents with very specific focus for code review and rule enforcement mostly sharing cache.", "context": "", "use_case": "agent fleet for code review", "switched_from": "", "story": "Boris Cherny says \"delete your Claude.md every 6 months\""}, {"id": 783711504, "item_id": 48981969, "url": "https://news.ycombinator.com/item?id=48981969", "by": "arikrahman", "date": "2026-07-20", "stance": "recommend", "quote": "I use the reasonix or whale harness for Deepseek, prefer the hyper optimized cache hits making request converge on almost free.", "excerpt": "This is why I use the reasonix or whale harness for Deepseek, prefer the hyper optimized cache hits making request converge on almost free.", "context": "", "use_case": "DeepSeek harness", "switched_from": "", "story": "Stop Using OpenCode"}]}, {"label": "Large contexts", "summary": "Cheap cached tokens make large-scale code analyses with mega context windows practical.", "quotes": [{"id": 797043538, "item_id": 49815221, "url": "https://news.ycombinator.com/item?id=49815221", "by": "abalashov", "date": "2026-09-23", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly", "excerpt": "Data points from my own usage, for whatever they're worth[1]. I'm just an individual contributor and hardly represent enterprise or large orgs: I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! 1) For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit i", "context": "individual contributor", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "The current balance of power in open models"}, {"id": 796953682, "item_id": 49809605, "url": "https://news.ycombinator.com/item?id=49809605", "by": "abalashov", "date": "2026-09-22", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache.", "excerpt": "Welcome to the dark side. I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit input tokens, so very cheap large-scale code analyses and things that require mega context windows (at the cost of some at", "context": "", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis"}]}, {"label": "Agent fleets", "summary": "One commenter runs many focused review and rule-enforcement agents that mostly share cache.", "quotes": [{"id": 785488995, "item_id": 49093062, "url": "https://news.ycombinator.com/item?id=49093062", "by": "wilj", "date": "2026-07-28", "stance": "recommend", "quote": "And Reasonix + Deepseek is mostly cache hits when I let it grind on it for a long time", "excerpt": "Similar here. I use Claude Code and Codex for developing specs, but then have Deepseek v4 Pro implement because it just seems to actually do what it's told, given enough tries. And Reasonix + Deepseek is mostly cache hits when I let it grind on it for a long time, and I can have a huge fleet of agents with very specific focus for code review and rule enforcement mostly sharing cache.", "context": "", "use_case": "agent fleet for code review", "switched_from": "", "story": "Boris Cherny says \"delete your Claude.md every 6 months\""}]}, {"label": "Planning workflow", "summary": "Paired with planning-with-files, one commenter found it better than openspec.", "quotes": [{"id": 780316016, "item_id": 48769751, "url": "https://news.ycombinator.com/item?id=48769751", "by": "arikrahman", "date": "2026-07-02", "stance": "recommend", "quote": "I use Reasonix with planning-with-files and that's been better than even openspec.", "excerpt": "Likewise. I use Reasonix with planning-with-files and that's been better than even openspec. Cache hits with the help of Reasonix on Deepseek make requests practically free. And that's unsubsidized, with American providers like Digital Ocean or cloudflare.", "context": "", "use_case": "AI coding with DeepSeek", "switched_from": "", "story": "Superpowers 6"}]}], "caveats": [{"label": "Data to China", "summary": "Using Reasonix requires sending data to China.", "quotes": [{"id": 796953682, "item_id": 49809605, "url": "https://news.ycombinator.com/item?id=49809605", "by": "abalashov", "date": "2026-09-22", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache.", "excerpt": "Welcome to the dark side. I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit input tokens, so very cheap large-scale code analyses and things that require mega context windows (at the cost of some at", "context": "", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis"}]}, {"label": "Attentional drift", "summary": "Very large context runs come at the cost of some attentional drift.", "quotes": [{"id": 797043538, "item_id": 49815221, "url": "https://news.ycombinator.com/item?id=49815221", "by": "abalashov", "date": "2026-09-23", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly", "excerpt": "Data points from my own usage, for whatever they're worth[1]. I'm just an individual contributor and hardly represent enterprise or large orgs: I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! 1) For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit i", "context": "individual contributor", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "The current balance of power in open models"}, {"id": 796953682, "item_id": 49809605, "url": "https://news.ycombinator.com/item?id=49809605", "by": "abalashov", "date": "2026-09-22", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache.", "excerpt": "Welcome to the dark side. I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit input tokens, so very cheap large-scale code analyses and things that require mega context windows (at the cost of some at", "context": "", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis"}]}, {"label": "OpenRouter caching", "summary": "A commenter questioned whether cache benefits carry over when routing DeepSeek through OpenRouter.", "quotes": [{"id": 772126704, "item_id": 48257919, "url": "https://news.ycombinator.com/item?id=48257919", "by": "schaefer", "date": "2026-05-24", "stance": "neutral", "quote": "If I use Reasonix to talk to a deepseek endpoint through openrouter", "excerpt": "Okay, I'm curious. From the FAQ, I see: >Can I point it at a self-hosted / private DeepSeek endpoint? >Yes. Since 0.30 we accept non-standard key prefixes for self-hosted DeepSeek endpoints. Just point `baseUrl` at your internal address \u2014 the loop, cache strategy, and tool protocol are unchanged. But my question is: If I use Reasonix to talk to a deepseek endpoint through openrouter, am I still getting the cache-hit ", "context": "", "use_case": "DeepSeek agent harness", "switched_from": "", "story": "DeepSeek reasonix, engineered for low token cost during long running sessions"}]}], "who": [{"by": "abalashov", "context": "individual contributor", "url": "https://news.ycombinator.com/item?id=49815221", "date": "2026-09-23"}], "tags": ["for DeepSeek users", "for large codebase analysis", "for multi-agent code review", "self-hosted endpoints", "for individual developers"], "use_cases": ["deepseek coding harness", "harness for deepseek", "agent harness", "coding agent harness", "agent fleet for code review", "deepseek harness"], "compared": [{"slug": "deepseek", "name": "DeepSeek", "co": 10, "switched_from": 0, "switched_to": 0}, {"slug": "openrouter", "name": "OpenRouter", "co": 5, "switched_from": 0, "switched_to": 0}, {"slug": "kimi", "name": "Kimi", "co": 5, "switched_from": 0, "switched_to": 0}, {"slug": "qwen-3-6", "name": "Qwen 3.6", "co": 5, "switched_from": 0, "switched_to": 0}, {"slug": "mimo", "name": "MiMo", "co": 4, "switched_from": 0, "switched_to": 0}, {"slug": "glm", "name": "GLM", "co": 4, "switched_from": 0, "switched_to": 0}, {"slug": "opencode", "name": "OpenCode", "co": 3, "switched_from": 0, "switched_to": 0}, {"slug": "pi", "name": "Pi", "co": 3, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=49808816", "title": "The current balance of power in open models", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49796660", "title": "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49672510", "title": "We must pace the frontier", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49273478", "title": "Qwen/Qwen3.8-2.4T-A95B", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49120299", "title": "DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49077040", "title": "Boris Cherny says \"delete your Claude.md every 6 months\"", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {}, "2026-02": {}, "2026-03": {}, "2026-04": {}, "2026-05": {"neutral": 1}, "2026-06": {"neutral": 1}, "2026-07": {"recommend": 5}, "2026-08": {"recommend": 1}, "2026-09": {"recommend": 1}, "2026-10": {}}, "quotes": [{"id": 797043538, "item_id": 49815221, "url": "https://news.ycombinator.com/item?id=49815221", "by": "abalashov", "date": "2026-09-23", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly", "excerpt": "Data points from my own usage, for whatever they're worth[1]. I'm just an individual contributor and hardly represent enterprise or large orgs: I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! 1) For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit i", "context": "individual contributor", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "The current balance of power in open models"}, {"id": 796953682, "item_id": 49809605, "url": "https://news.ycombinator.com/item?id=49809605", "by": "abalashov", "date": "2026-09-22", "stance": "recommend", "quote": "I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache.", "excerpt": "Welcome to the dark side. I've been on Kimi, with a little DeepSeek-V4-Pro, GLM 5.2/5.3, and MiMo thrown in, for probably about a year now. It's great here! For DeepSeek, I recommend their Reasonix harness strongly, due to its alignment to DeepSeek's prefix cache. It means mostly (95%+) cache hit input tokens, so very cheap large-scale code analyses and things that require mega context windows (at the cost of some at", "context": "", "use_case": "DeepSeek coding harness", "switched_from": "", "story": "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis"}, {"id": 794800339, "item_id": 49675021, "url": "https://news.ycombinator.com/item?id=49675021", "by": "abalashov", "date": "2026-09-12", "stance": "recommend", "quote": "DeepSeek via the Reasonix harness", "excerpt": "Allow me to recommend switching to Kimi K3 / GLM 5.3 / DeepSeek. I use all three for the better part of a year now (DeepSeek via the Reasonix harness) and haven't touched Claude Code in at least as long.", "context": "", "use_case": "harness for DeepSeek", "switched_from": "", "story": "We must pace the frontier"}, {"id": 788624931, "item_id": 49289058, "url": "https://news.ycombinator.com/item?id=49289058", "by": "SwellJoe", "date": "2026-08-13", "stance": "recommend", "quote": "I also like Reasonix quite a bit, as an agent harness", "excerpt": "DeepSeek is more fun, because it's cheap-as-free, Good Enough, quite fast. I use it for all API stuff, automated runs, testing of the security auditing harness and benchmarks I'm working on, etc. Kimi K3 is smarter, though. At least smarter than DeepSeek V4 Flash 0731. I haven't tried the new Pro version, but will this weekend when I'm working on my personal projects But, K3 has been what I've been using for the actu", "context": "", "use_case": "agent harness", "switched_from": "", "story": "Qwen/Qwen3.8-2.4T-A95B"}, {"id": 785986657, "item_id": 49124166, "url": "https://news.ycombinator.com/item?id=49124166", "by": "bwfan123", "date": "2026-07-31", "stance": "recommend", "quote": "With reasonix or pi, I can code all day long and pay a few pennies for it.", "excerpt": "> For the Code Agent tasks among the public benchmarks above, DeepSeek-V4-Flash-0731 is evaluated with the minimal mode of DeepSeek Harness (to be released) as the agent framework So, are they planning to announce an optimized coding agent harness as well ? DSv4 flash is a fantastic model, and my daily driver. With reasonix or pi, I can code all day long and pay a few pennies for it. No token anxiety. Whereas the sam", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis"}, {"id": 785488995, "item_id": 49093062, "url": "https://news.ycombinator.com/item?id=49093062", "by": "wilj", "date": "2026-07-28", "stance": "recommend", "quote": "And Reasonix + Deepseek is mostly cache hits when I let it grind on it for a long time", "excerpt": "Similar here. I use Claude Code and Codex for developing specs, but then have Deepseek v4 Pro implement because it just seems to actually do what it's told, given enough tries. And Reasonix + Deepseek is mostly cache hits when I let it grind on it for a long time, and I can have a huge fleet of agents with very specific focus for code review and rule enforcement mostly sharing cache.", "context": "", "use_case": "agent fleet for code review", "switched_from": "", "story": "Boris Cherny says \"delete your Claude.md every 6 months\""}, {"id": 783711504, "item_id": 48981969, "url": "https://news.ycombinator.com/item?id=48981969", "by": "arikrahman", "date": "2026-07-20", "stance": "recommend", "quote": "I use the reasonix or whale harness for Deepseek, prefer the hyper optimized cache hits making request converge on almost free.", "excerpt": "This is why I use the reasonix or whale harness for Deepseek, prefer the hyper optimized cache hits making request converge on almost free.", "context": "", "use_case": "DeepSeek harness", "switched_from": "", "story": "Stop Using OpenCode"}, {"id": 783536705, "item_id": 48971044, "url": "https://news.ycombinator.com/item?id=48971044", "by": "abalashov", "date": "2026-07-19", "stance": "recommend", "quote": "When I want a harness to inhale a codebase and give me an answer, Reasonix has been the go-to for a while now, and it's great.", "excerpt": "I just pay metered usage via OpenRouter. I've spent no more than $10-$20/mo, but my usage is, by the standards of so-called \"agentic coding\" these days, very light and confined. I doubt my workflows match the profile of \"daily driver\" as you're conceiving of it. I'm mostly in the IDE chat box asking for a hand here and there, but no \"long horizon\" / big vision / architecture things. Occasionally, I'll solicit a gener", "context": "", "use_case": "codebase analysis", "switched_from": "", "story": "Moonshot AI suspends new subscriptions due to Kimi K3 demand"}, {"id": 783398945, "item_id": 48962434, "url": "https://news.ycombinator.com/item?id=48962434", "by": "SwellJoe", "date": "2026-07-18", "stance": "recommend", "quote": "with Reasonix, DeepSeek is free real estate. Seems to just go and go for pennies.", "excerpt": "Yep, with Reasonix, DeepSeek is free real estate. Seems to just go and go for pennies. And, DeepSeek is what I use for any task that works best with an API. It's cheap enough to where I don't think about cost, made even cheaper by DeepSeek having the most effective and cheap caching in the industry, and it's good enough to where I rarely have to follow up with a more expensive model or manually fix things. It's been ", "context": "", "use_case": "running DeepSeek cheaply", "switched_from": "", "story": "The Kimi K3 Moment"}, {"id": 780471571, "item_id": 48779473, "url": "https://news.ycombinator.com/item?id=48779473", "by": "SwellJoe", "date": "2026-07-03", "stance": "recommend", "quote": "especially with Reasonix agent (which maximizes caching, which DeepSeek is uniquely good at", "excerpt": "I recently wrote up how I run local LLMs, because several folks had asked (https://swelljoe.com/post/how-i-run-local-llms/) and I think even my setup, which I spent maybe $4200 on, half on a Strix Halo and half on upgrades for my desktop, would be too expensive to justify today. I bought before prices went through the roof, and only did so because I like to tinker with hardware...not because I expected it to ever pay", "context": "", "use_case": "DeepSeek coding agent", "switched_from": "", "story": "Jamesob's guide to running SOTA LLMs locally"}, {"id": 780316016, "item_id": 48769751, "url": "https://news.ycombinator.com/item?id=48769751", "by": "arikrahman", "date": "2026-07-02", "stance": "recommend", "quote": "I use Reasonix with planning-with-files and that's been better than even openspec.", "excerpt": "Likewise. I use Reasonix with planning-with-files and that's been better than even openspec. Cache hits with the help of Reasonix on Deepseek make requests practically free. And that's unsubsidized, with American providers like Digital Ocean or cloudflare.", "context": "", "use_case": "AI coding with DeepSeek", "switched_from": "", "story": "Superpowers 6"}, {"id": 780015571, "item_id": 48750973, "url": "https://news.ycombinator.com/item?id=48750973", "by": "SwellJoe", "date": "2026-07-01", "stance": "recommend", "quote": "Reasonix is an excellent CLI harness that is designed around maximizing cacheability  of DeepSeek models", "excerpt": "North Mini Code is on par with the similar-sized MoE Qwen 3.6 and Gemma 4. It benches a little better than Gemma 4 26b a4b and a little worse than Qwen 3.6 35b a3b for most coding-related and agentic tasks. For my use case, or at least the use case I've tested, it is a little worse than both for surfacing security bugs, but all of the MoE models at this size are pretty bad at finding security bugs (they hallucinate a", "context": "", "use_case": "DeepSeek CLI harness", "switched_from": "", "story": "Qwen 3.6 27B is the sweet spot for local development"}, {"id": 776934659, "item_id": 48558416, "url": "https://news.ycombinator.com/item?id=48558416", "by": "abalashov", "date": "2026-06-16", "stance": "neutral", "quote": "DeepSeek through OpenRouter and Reasonix", "excerpt": "And if you want to dial in a setting in between: I've switched to Kimi K2.6 (now K2.7) and DeepSeek through OpenRouter and Reasonix for pretty much everything, with no discernible loss of analytical quality or utility. However, like many commenters, I don't really believe in vibe-coding, long-horizon agentic one-shot agentic coding, etc. and do not use LLMs for huge generation tasks that involve designing things end-", "context": "", "use_case": "", "switched_from": "", "story": "Running local models is good now"}, {"id": 772126704, "item_id": 48257919, "url": "https://news.ycombinator.com/item?id=48257919", "by": "schaefer", "date": "2026-05-24", "stance": "neutral", "quote": "If I use Reasonix to talk to a deepseek endpoint through openrouter", "excerpt": "Okay, I'm curious. From the FAQ, I see: >Can I point it at a self-hosted / private DeepSeek endpoint? >Yes. Since 0.30 we accept non-standard key prefixes for self-hosted DeepSeek endpoints. Just point `baseUrl` at your internal address \u2014 the loop, cache strategy, and tool protocol are unchanged. But my question is: If I use Reasonix to talk to a deepseek endpoint through openrouter, am I still getting the cache-hit ", "context": "", "use_case": "DeepSeek agent harness", "switched_from": "", "story": "DeepSeek reasonix, engineered for low token cost during long running sessions"}]}