AI design: agents × models on one taskhttps://mikeozornin.ru/recipe-scaler-bench/en/enhttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-gemini-4-argon-high/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-gemini-4-argon-high/Very AI-sloppy overall. Feels trained on B2B products — packed with everything, built a spaceship, wildly overloaded everywhere. The “interface” concept is generally clear, though. The promo page is very bad. (Mid)Wed, 14 Oct 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-cat-jam/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-cat-jam/The result is weak, kind of AI-slop-ish. But it was so fast. 10M tokens on top of that (obviously cache reads, but everyone's are cache reads). It must be a very small model. Would be interesting to know what this stealth model is. (Weak)Tue, 13 Oct 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-mistral-large-4-high/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-mistral-large-4-high/The best of the country's top-3 models, but it can't really catch the top-2 and top-1 models. Weak all around. (Weak)Fri, 09 Oct 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-fable-5-1-xhigh/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-fable-5-1-xhigh/The promo page is weak, but mobile and desktop are very well structured and good (desktop especially). Honestly, not shockingly expensive (but it did cut corners on the promo page). (Fancy)Thu, 08 Oct 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-haiku-5-5-max/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-haiku-5-5-max/Very slow, but cheap. Understood the desktop app concept well: it really is an app. Desktop and mobile are usable overall. Visually not good enough anywhere. The copy is much weaker than the bigger models'. (Mid)Thu, 08 Oct 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-sonnet-5-5-max/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-sonnet-5-5-max/Incredibly slow, but it's max. Very good. Pleasant to look at, interesting to watch. (Fancy)Thu, 08 Oct 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-6-1-sol-xhigh/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-6-1-sol-xhigh/Good copy, on a schoolmate level. No real design for a model of this class: the desktop is bad, the promo page is uninteresting. (Mid)Thu, 08 Oct 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-step-5-preview-free/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-step-5-preview-free/A typical middling run (middling, minus): formally everything is done, but there is nothing to hold on to — plain layout, the desktop is not great, the copy is weak. (Mid)Thu, 08 Oct 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-ling-3-1-flash/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-ling-3-1-flash/Nothing interesting overall; it would be mid if it had handled autolayout. Mobile is well structured, the rest is worse. (Weak)Fri, 02 Oct 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cline-pixel-canary/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cline-pixel-canary/Couldn't wait for it to finish — showing the result after 1 h 11 min. Only the promo page of the three screens was done. (Weak)Sun, 27 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-longcat-2-5-preview/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-longcat-2-5-preview/Bad, failed at everything. Tokens are likely reported without cache — can't be that low. (Failed tools)Fri, 25 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-longcat-2-0/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-longcat-2-0/All bad, failed with the tools. Tokens may be counted without cache — unclear. (Failed tools)Fri, 25 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-gemini-3-8-flash/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-gemini-3-8-flash/On par with Chinese models from six months ago. Weak across the board: visuals, logic, copy. Ran via OpenRouter proxy with thinking disabled (0 reasoning tokens), gemini-3.8-flash-20260902 snapshot. (Weak)Thu, 24 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-space-bunny-alpha/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-space-bunny-alpha/Weak across the board: layout, the "app" idea, the copy. (Weak)Thu, 24 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-6-luna/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-6-luna/For Luna's price — a delight. Overall not meaningfully worse than Astra. Missed the desktop app and made some editorial site. Copywriting is weak. (Mid)Wed, 23 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-6-sol/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-6-sol/Got the desktop logic right: made an actual app, not a catalog site. Everything needed is there. Visually weak, promo page is weak. Fast and token-efficient, so cheap. (Mid)Wed, 23 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-opus-5-5/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-opus-5-5/Really strong: right logic and accents, solid layout, visually precise, good copy (finally!). Looks like work by a very decent designer. There are small flaws, but I'm impressed. (Fancy)Wed, 23 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-hy4-preview/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-hy4-preview/Very bad — failed the tools so hard it burned through the 100-call free-tier MCP limit. Models usually get by on half of that. (Failed tools)Tue, 22 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-mimo-v2-6-pro/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-mimo-v2-6-pro/Ordinary, nothing special to grab onto. Low-mid tier. (Mid)Tue, 22 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-grok-4-7/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-grok-4-7/The result is weak for a model of this level and its release date. Very efficient on tokens. (Mid)Mon, 21 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-grok-4-6-2/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-grok-4-6-2/Still a bit weak for a model of this level. Colors and style somehow like GPT 5.4. Cursor didn't count tokens for some reason. (Mid)Sun, 20 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-kimi-2-8/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-kimi-2-8/Weak at the high end. Upside: comparable to K3. (Weak)Mon, 14 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/figma-claude-code-opus-5/https://mikeozornin.ru/recipe-scaler-bench/en/run/figma-claude-code-opus-5/Typical Anthropic and Opus (Fancy)Fri, 11 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-5/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-5/Typical Opus, very similar to previous generations. But the color is unusual — where’s the beige, Opus? :–) (Fancy)Fri, 11 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-deepseek-4-1-flash/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-deepseek-4-1-flash/Slow, partly because the model did not read the images itself and instead reached for the available zai MCP server to analyse them. That would make sense for regular DeepSeek, but this one has vision. Lots of re-checking, as if I had typed /goal make the most fantastic design ever. (Mid)Thu, 10 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-6-astra/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-6-astra/Really strong copywriting (hadn't paid attention to it before). Design is average, not SOTA. (Mid)Sat, 05 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-muse-spark-1-3-free/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-muse-spark-1-3-free/Nice, big progress for the model. Made very fast. * Product by Meta, recognized as extremist and banned in Russia. (Fancy)Fri, 04 Sep 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-glm-5-3-flash/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-glm-5-3-flash/Much weaker than full GLM 5.3 Max: vision is there, but layout bugs (servings overlap, clipping) were not fixed. ~9.7M tokens in 21 min. (Weak)Sun, 30 Aug 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-glm-5-3-max/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-glm-5-3-max/Clearly much better than GLM 5.2, which couldn't even use tools properly for me (will rerun it). Spent the first 10 minutes studying images via zai MCP; the model isn't multimodal. (Mid)Sat, 15 Aug 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-grok-4-6/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-grok-4-6/Fast, overall decent, very distinctive. Pros: didn't turn the desktop UI into a promo site. Cons: still not quite great — could have put images somewhere (Mid)Wed, 12 Aug 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-kimi-k3/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-kimi-k3/— (Weak)Fri, 17 Jul 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/figma-claude-code-opus-4-7/https://mikeozornin.ru/recipe-scaler-bench/en/run/figma-claude-code-opus-4-7/Opus in its style, especially from afar, but ate more than a full 5-hour quota (Fancy)Tue, 14 Jul 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/figma-codex-gpt-5-5/https://mikeozornin.ru/recipe-scaler-bench/en/run/figma-codex-gpt-5-5/In Figma slightly better than in Paper. Lays things out monstrously simply (Weak)Tue, 14 Jul 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/figma-cursor-composer-2/https://mikeozornin.ru/recipe-scaler-bench/en/run/figma-cursor-composer-2/Catastrophically bad, worse than in Paper (Weak)Tue, 14 Jul 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/figma-cursor-glm-5-1/https://mikeozornin.ru/recipe-scaler-bench/en/run/figma-cursor-glm-5-1/In general worse than in Paper. GLM didn't master a correct DOM (Weak)Tue, 14 Jul 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-deepseek-v4-pro-2/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-deepseek-v4-pro-2/DeepSeek is very cheap, but managed only once. Mobile is watchable (Failed tools)Tue, 14 Jul 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-5-6-sol/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-5-6-sol/— (Fancy)Mon, 13 Jul 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-grok-4-5/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-grok-4-5/Very fast, cheap (50% discount for now), overall fine for the money. Promo is primitive (Mid)Wed, 08 Jul 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-fable-med/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-fable-med/I thought it would burn more than it did. Overall little distinguishable from Opus, in places even simpler. (Fancy)Thu, 11 Jun 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-code-nemotron-3-ultra/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-code-nemotron-3-ultra/Weak, nothing outstanding (Weak)Mon, 01 Jun 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-minimax-m3/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-minimax-m3/M2.7 → M3 — a huge step. From flop into top 3. Very similar in style to Opus (Fancy)Mon, 01 Jun 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-4-8/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-4-8/Burns more; not necessarily better than 4.7. Three screens used half of the already-doubled 5h pro quota. (Fancy)Thu, 28 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-mimo-v2-5-pro/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-mimo-v2-5-pro/Included Cursor model; very cheap in subscription terms, weak on design (Weak)Wed, 27 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-code-grok-build-google-skill/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-code-grok-build-google-skill/Same prompt with Google's design skill; compare with OpenCode Grok Build 0.1 without a skill (Weak)Mon, 25 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-grok-build-0-1/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-grok-build-0-1/Fast, weak, expensive for this quality (Weak)Fri, 22 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-qwen-3-7/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-qwen-3-7/Expensive for its result, no improvement over 3.6 Max. Very bad at using tools (Weak)Thu, 21 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-antigravity-gemini-3-5-flash/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-antigravity-gemini-3-5-flash/Noticeably better than Gemini 3.1 Pro. The only one who decided to do a dark mode. (Mid)Wed, 20 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-composer-2-5/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-composer-2-5/Comparable in cost to Composer 2, quality has grown, but still very not great (Weak)Wed, 20 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-hy3-preview/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-hy3-preview/The trendiest open-source model on Openrouter. Very so-so (Weak)Mon, 11 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-sourcecraft-default/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-sourcecraft-default/Seems inexpensive, but no point in using it, neuro-slop (Weak)Mon, 11 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-sourcecraft-default-thinking/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-sourcecraft-default-thinking/Default and Default Thinking seem like different models; reasoning is even worse (Weak)Mon, 11 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-4-6/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-4-6/In general looks like a simplified Opus 4.7, much closer to Sonnet in overall layout (Mid)Sun, 10 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-frontend-design/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-frontend-design/Opus with Frontend Design skill — desktop only in local asset set (Fancy)Thu, 07 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-auto/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-auto/A surprisingly good result for Auto mode. The style choice is like from Opus (Mid)Thu, 07 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-antigravity-gemini-3-1-pro/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-antigravity-gemini-3-1-pro/Very bad at using tools, monstrous integration into the tool (Weak)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-haiku-4-5/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-haiku-4-5/Didn't even cope with the tools. There's not a single reason to use it (Failed tools)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-4-7/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-4-7/Quick. Expensive. Awesome. (Fancy)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-4-7-med/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-opus-4-7-med/Keeps the Opus 4.7 style, simplifies the implementation (Fancy)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-sonnet-4-5/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-claude-code-sonnet-4-5/Substantially closer to the Chinese models and the simpler ones. Neat, but completely neutral, completely simplified (Mid)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-5-4/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-5-4/Surprisingly, I like this variant even more than GPT 5.5 (Mid)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-5-5/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-codex-gpt-5-5/Very cheap compared to opus, the result is accordingly. The worst result among SOTA models (Weak)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-composer-2/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-composer-2/Very primitive, but also very cheap. Doesn't match the claimed level (Weak)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-gemini-3-1-pro/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-gemini-3-1-pro/Result comparable to Gemini in Antigravity. Very bad for a model of this level (Weak)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-glm-5-1/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-glm-5-1/GLM in Cursor for some reason couldn't cope with the tools (Failed tools)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-gpt-5-4/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-gpt-5-4/The agent is different, the model is recognizable. Slightly better than in Codex (Mid)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-grok-4-3/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-grok-4-3/Mixed impressions. Mobile is generally fine. Promo is worse (Weak)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-opus-4-7/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-cursor-opus-4-7/Twice as expensive as Opus 4.7 at Kilo. Checks itself a lot. The result is slightly better (Fancy)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-code-opus-4-7/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-code-opus-4-7/Typical Opus 4.7. Clean in places, interesting, from afar it's super-duper (Fancy)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-code-qwen-3-5-397/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-code-qwen-3-5-397/Didn't cope with the tools (Failed tools)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-code-qwen-3-6/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-kilo-code-qwen-3-6/In general Qwen did decently. A bit dim, but pretty solid. 1/30 of Opus's price (Mid)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-deepseek-v4-pro/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-deepseek-v4-pro/DeepSeek is very cheap, but managed only once. Mobile is watchable (Failed tools)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-minimax-2-7/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-minimax-2-7/Incredibly fast, incredibly cheap, the quality is accordingly (Weak)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-qwen-3-5-397/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-qwen-3-5-397/Couldn't do anything (Failed tools)Wed, 06 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-glm-5-1/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-glm-5-1/Unlike Cursor, Opencode got something out of GLM — OK for its price (Mid)Tue, 05 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-grok-4-3/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-grok-4-3/Unlike Cursor, Opencode couldn't handle the model at all (Weak)Tue, 05 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-kimi-2-6/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-kimi-2-6/Mobile is better than the rest. Everything was duplicated, then erased via screenshots — 40+ minutes (Weak)Tue, 05 May 2026 12:00:00 GMThttps://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-opus-4-7/https://mikeozornin.ru/recipe-scaler-bench/en/run/paper-opencode-opus-4-7/Not bad; after this screen Openrouter banned American SOTA models (Fancy)Tue, 05 May 2026 12:00:00 GMT