Partial day — headline is the last complete day (2026-10-13).
twitter
tracked since 2026-10-11T14:10:50.577329+00:00
conclusion
Measured mentions per day
day (UTC)
mentions
2026-10-04
1
2026-10-05
3
2026-10-06
0
2026-10-07
11
2026-10-08
6
2026-10-09
5
2026-10-10
8
2026-10-11
5
2026-10-12
5
2026-10-13
2
2026-10-14
0
derived from
haiku
word-boundary match over title + text, query='timeline' · window
2026-10-04 → 2026-10-15
· 46 matching posts
evidenceShowing 20 newest of 46
•
twitterrvaniaaaa20hpositive
THIS IS F*CKING BRILLIANT. Anthropic’s docs describe one of the smartest ways to run Opus
three models work on one task and each does only the part it’s cheapest at. start Claude Code like this:
claude -> model sonnet -> advisor opus
you stop paying Opus prices for work that doesn’t need Opus
> Sonnet runs the sess…THIS IS F*CKING BRILLIANT. Anthropic’s docs describe one of the smartest ways to run Opus
three models work on one task and each does only the part it’s cheapest at. start Claude Code like this:
claude -> model sonnet -> advisor opus
you stop paying Opus prices for work that doesn’t need Opus
> Sonnet runs the session and edits the code
> Haiku 5.5 searches the repo and docs, at up to 40x less per token than Opus 5.5
> Opus waits as the advisor and reads the whole session, tool calls included
Claude usually calls Opus at three points:
> before it locks a plan
> when the same error comes back
> right before it says it’s done
Sonnet drives. Haiku scouts. Opus signs off
and if you go one step lower, most of an agent’s day is tiny choices like picking a file or a tool, or deciding whether to retry. Jev takes those first, so Sonnet and Opus only see the moments that are really open
the whole setup
Sonnet runs the main session
Explore on Haiku 5.5 searches and pulls docs, read-only
Opus on call as the advisor
give Claude Code this setup and the prompt below ↓
“Set up my Claude Code exactly like this:
Main model Sonnet, advisor Opus, saved as advisorModel in my settings.
One read-only subagent named exactly Explore, on claude-haiku-5-5, to replace the built-in one.
One line in CLAUDE.md: ask the advisor before locking a plan, after the same error twice, and before calling a task done, not more often.
Check what could turn the advisor off (CLAUDE_CODE_DISABLE_ADVISOR_TOOL, DISABLE_TELEMETRY) and only report it.
Show every change as a diff first. Don’t edit anything until I say go.”
https://t.co/xpEE0hFQ6d
before you turn it on:
every advisor call re-reads the whole transcript with no cache. on a 200k session that’s about $0.80 in input, against $0.04 for a normal request. keep it to a few calls
the built-in Explore agent runs on your main model unless a subagent is named exactly Explore. any other name does nothing
turning the advisor on or off keeps y
Muse Wonder & Wonder FULL by @AIatMeta on KingBench 4
Ranking:
Opus 5.5 — 57.8%
GPT-6 Astra — 56.5%
GPT-6.1 Sol — 55.0%
Haiku 5.5 — 54.2%
Muse Wonder Full 1.0 — 53.0%
Muse Wonder 1.0 and GLM-5.3 — 45.8% each, tied
The models is not Opus 5.5 Level and is not token efficient at all. There are more downsides to the mode…Muse Wonder & Wonder FULL by @AIatMeta on KingBench 4
Ranking:
Opus 5.5 — 57.8%
GPT-6 Astra — 56.5%
GPT-6.1 Sol — 55.0%
Haiku 5.5 — 54.2%
Muse Wonder Full 1.0 — 53.0%
Muse Wonder 1.0 and GLM-5.3 — 45.8% each, tied
The models is not Opus 5.5 Level and is not token efficient at all. There are more downsides to the model than there are upsides.
GPT-6.1 Sol still remains the king at being token efficient.
Sol costs about $1 per task,
Haiku costs about $1.23 per task,
Wonder costs about $1.75 per task.
Wonder Full costs about $2.87 per task
Astra costs $6.6 and Opus 5.5 - $7.5
This price could have been very justified if the performance was on level with Opus or Astra but it is below Sol and Haiku. Wonder is on GLM-5.3 level while Wonderfull is short of Sol & Haiku.
Wonder is really bad at complex prompts, developing fully featured apps unattended.
The bench for context tests for 6 prompts that involve building a docker clone (VM), GBA, DS Emulation, Room Desigining Software, Pokemon Red reconstruction, QMK/VIA Compatible Keyboard Configurator and so on. The model is not acing any of these tests. The full test results will be available on the YT Video: https://t.co/Z1XMPnAbIL
Huge model launch:
Claude Haiku 5.5 is free forever in our coding agent 🥳
- Get 2 hours per day of Haiku in the US
- Funded by ads, so it's sustainable
Try it now: https://t.co/uCHw0PuzKO https://t.co/u4aZkx0rJnHuge model launch:
Claude Haiku 5.5 is free forever in our coding agent 🥳
- Get 2 hours per day of Haiku in the US
- Funded by ads, so it's sustainable
Try it now: https://t.co/uCHw0PuzKO https://t.co/u4aZkx0rJn
THIS IS ACTUAL F*CKING GOLD
Anthropic just dropped one of the best tricks for Claude Code.
The idea: run Sonnet, Haiku and Opus 5.5 together, each one doing only the job it's best at.
Better output and way fewer tokens, because every task lands on the right model.
You just launch Claude Code like this:
claude --mo…THIS IS ACTUAL F*CKING GOLD
Anthropic just dropped one of the best tricks for Claude Code.
The idea: run Sonnet, Haiku and Opus 5.5 together, each one doing only the job it's best at.
Better output and way fewer tokens, because every task lands on the right model.
You just launch Claude Code like this:
claude --model sonnet --advisor opus
here's the split:
→ Sonnet 5 runs the session and writes the code
→ Haiku 5.5 explores files and docs in parallel
→ Opus 5.5 stays on call as advisor for the big calls
→ you stop burning Opus context on work that doesn't need it
Opus only steps in at three moments:
→ before a plan locks
→ when the same error shows up twice
→ before calling it done
Sonnet builds. Haiku explores. Opus reviews.
JEV engineering is the same move one layer down: the forks that need no thinker (which file, which tool, retry or stop) go to JEV in milliseconds, and the big models only see the ones that split
the full tree
> Sonnet 5 runs the main session and edits code
> search maps the repo on Haiku 5.5
> grep finds call sites on Haiku 5.5
> docs pulls the API reference on Haiku 5.5
> Opus 5.5 on call as advisor, reads the whole transcript when consulted
> JEV routes micro-forks in milliseconds
paste the tree and this prompt into Claude Code ↓
"Rebuild my Claude Code setup around this tree:
1. Run the main session on Sonnet 5 with Opus 5.5 as advisor.
2. Create three read-only Haiku 5.5 subagents: search (maps the repo), grep (finds call sites), docs (pulls the API reference).
3. Add one rule to CLAUDE.md: consult the advisor before locking a plan, after the same error twice, and before calling a task done.
4. Route fast deterministic forks (tool choice, path selection, retries) through JEV.
Show me every change as a diff first. No edits until I say go."
↳ https://t.co/vOifNVjYxB
you'll use this every single day.
Your most expensive model shouldn't be doing your cheapest work.
This Claude Code workflow splits the job across Opus 5.5, Sonnet 5.5, and Haiku 5.5:
→ Opus plans and synthesizes the final PR
→ Sonnet handles precise refactors in isolated worktrees
→ Haiku filters candidates at 340 tok/s
The clever part? Haiku steps …Your most expensive model shouldn't be doing your cheapest work.
This Claude Code workflow splits the job across Opus 5.5, Sonnet 5.5, and Haiku 5.5:
→ Opus plans and synthesizes the final PR
→ Sonnet handles precise refactors in isolated worktrees
→ Haiku filters candidates at 340 tok/s
The clever part? Haiku steps in at 3 critical bottlenecks:
→ Before dispatch: Scans the codebase in parallel and extracts AST candidates without bloating context.
→ During validation: Catches malformed tool calls and false positives before Sonnet touches them.
→ Before synthesis: Checks that every candidate matches the target JSON schema.
Haiku filters. Sonnet refactors. Opus ships.
Here's how to set up the full workflow in Claude Code:
Let Opus plan. Use high effort to draft workflow scripts in .claude/workflows/.
Fan out Haiku. Run parallel AST sweeps across target files at low effort.
Validate candidates. Use schema checks to filter false positives before generating patches.
Let Sonnet refactor. Run verified changes in isolated Git worktrees to avoid edit conflicts.
Keep intermediate state out of context. Use runtime variables to avoid carrying unnecessary data between agents.
Cache prompt prefixes. Configure a 1-hour prompt cache TTL to reduce redundant processing where supported.
The same principle applies to agent frameworks like Jev: use deterministic code for simple routing decisions, and call frontier models only when reasoning is actually needed.
Author on high. Refactor on medium. Let Haiku handle the filtering.
Want to try it? Paste this prompt into Claude Code:
Configure my Claude Code workspace for dynamic workflows.
1. Audit ~/.claude/workflows and .claude/workflows.
- Draft workflows for codebase audits, parallel migrations, and schema validation.
- Use Haiku at low effort for AST sweeps.
- Use Sonnet at medium effort for refactoring.
- Run workers in isolated Git worktrees.
2. Review ~/.claude/settings.json for supported workflow settin
THE BEST AI AGENT IN 2026 USES 4 MODELS.
Not one.
Opus 5.5 for the hard thinking.
Sonnet 5.5 for the building.
Haiku 5.5 for the volume.
Jev for the yes or no decisions.
Picking which model does which job used to take an engineer.
Now Opus 5.5 does it for you.
BTHE BEST AI AGENT IN 2026 USES 4 MODELS.
Not one.
Opus 5.5 for the hard thinking.
Sonnet 5.5 for the building.
Haiku 5.5 for the volume.
Jev for the yes or no decisions.
Picking which model does which job used to take an engineer.
Now Opus 5.5 does it for you.
B
Most people are using Claude Code wrong.
They're burning Opus tokens on work Haiku could handle in parallel.
Opus reviews. Sonnet builds. Haiku swarms.
Anthropic's multi-agent workflow lets you split the work:
→ Sonnet runs the main session, writes code, and executes tests
→ Haiku subagents search files, scan code,…Most people are using Claude Code wrong.
They're burning Opus tokens on work Haiku could handle in parallel.
Opus reviews. Sonnet builds. Haiku swarms.
Anthropic's multi-agent workflow lets you split the work:
→ Sonnet runs the main session, writes code, and executes tests
→ Haiku subagents search files, scan code, and fetch docs in parallel
→ Opus reviews architecture, catches regressions, and audits the final diff
the setup starts with one command:
`CLAUDE_CODE_SUBAGENT_MODEL=haiku claude --model sonnet`
Aa
Grid + MinGRU, 34% of the deck's 150M steps/s target, 10th of 18 (Opus 5.5: 79%, Opus 5 leads at 196%). The leaders fuse the rollout into one persistent kernel on fp16 tensor cores. Haiku uses four fp32 launches per step, but finds a closed-form jump for the shared food RNG.
Kernel: https://t.co/iWNaAf0Z1C
Trace: http…Grid + MinGRU, 34% of the deck's 150M steps/s target, 10th of 18 (Opus 5.5: 79%, Opus 5 leads at 196%). The leaders fuse the rollout into one persistent kernel on fp16 tensor cores. Haiku uses four fp32 launches per step, but finds a closed-form jump for the shared food RNG.
Kernel: https://t.co/iWNaAf0Z1C
Trace: https://t.co/QudOW0rD68
MegaQwen Decode, 4.6% of roofline, 11th of 21 (Opus 5.5: 7.4%, Sonnet 5.5: 8.0%). Opus and Sonnet run every decode step and all four layers in one cooperative launch. Haiku launches 29 small kernels per step from the host, so short contexts are launch-bound; at 128K it streams the cache as fast as the others.
Kernel: …MegaQwen Decode, 4.6% of roofline, 11th of 21 (Opus 5.5: 7.4%, Sonnet 5.5: 8.0%). Opus and Sonnet run every decode step and all four layers in one cooperative launch. Haiku launches 29 small kernels per step from the host, so short contexts are launch-bound; at 128K it streams the cache as fast as the others.
Kernel: https://t.co/aFaRmfYVbl
Trace: https://t.co/DevbUBGTeO
Fused MoE, 9.5% of roofline, 8th of 21 (Opus 5.5: 10.4%, Sonnet 5.5 leads at 11.3%). Same structure as both: sort tokens by expert, two grouped bf16 tensor-core GEMMs with SiLU-times-up fused into the first. Haiku skips the small-group tiling and fused combine, and runs the single-token decode shape through the same 12…Fused MoE, 9.5% of roofline, 8th of 21 (Opus 5.5: 10.4%, Sonnet 5.5 leads at 11.3%). Same structure as both: sort tokens by expert, two grouped bf16 tensor-core GEMMs with SiLU-times-up fused into the first. Haiku skips the small-group tiling and fused combine, and runs the single-token decode shape through the same 128-row tiles.
Board: https://t.co/BiHtR9p04C
Kernel: https://t.co/8T9P9SoYGn
Trace: https://t.co/YZ3fBJCo4m
Claude Haiku 5.5 [max] on KernelBench-CUDA:
Four hand-written CUDA kernels passed their final checks on RTX PRO 6000, all mid-board: 8th, 7th, 11th and 10th. Haiku picks the same plans as Opus 5.5 and Sonnet 5.5 but stops at the first working version: no tensor cores outside the MoE, and many small launches where the …Claude Haiku 5.5 [max] on KernelBench-CUDA:
Four hand-written CUDA kernels passed their final checks on RTX PRO 6000, all mid-board: 8th, 7th, 11th and 10th. Haiku picks the same plans as Opus 5.5 and Sonnet 5.5 but stops at the first working version: no tensor cores outside the MoE, and many small launches where the bigger models fuse into one.
On Mega it wrote no kernel. Its second session ended on Claude Code's 128K output-token cap.
One problem per reply below.
Guys, what AI model do you use for coming up with ideas?
I’m stuck between GPT 6.1 Sol and Claude Haiku 5.5. 🤔Guys, what AI model do you use for coming up with ideas?
I’m stuck between GPT 6.1 Sol and Claude Haiku 5.5. 🤔
holy sh*t. Official Anthropic tip for Claude Code: stop burning Opus 5.5 context on tasks Haiku 5.5 can swarm, while Sonnet 5.5 builds
put the architect on call with custom subagents
CLAUDE_CODE_SUBAGENT_MODEL=haiku claude --model sonnet
Sonnet 5.5 drives the primary session on high effort, writing diffs and running…holy sh*t. Official Anthropic tip for Claude Code: stop burning Opus 5.5 context on tasks Haiku 5.5 can swarm, while Sonnet 5.5 builds
put the architect on call with custom subagents
CLAUDE_CODE_SUBAGENT_MODEL=haiku claude --model sonnet
Sonnet 5.5 drives the primary session on high effort, writing diffs and running test suites
Haiku 5.5 subagents swarm the repo in parallel at 340 tok/s, handling file discovery and spec docs lookup
Opus 5.5 stays on call in ~/.claude/agents/ as the code reviewer, only stepping in at three critical moments:
→ before a plan locks: does the implementation plan miss auth invariants or schema contracts?
→ when a test breaks twice: are we fixing the root cause or falling into a recursive rabbit hole?
→ before calling done: did the full diff introduce hidden regressions or break pre-flight rules?
Opus 5.5 reviews. Sonnet 5.5 builds. Haiku 5.5 swarms
Claude Code subagent architecture runs the identical pattern through isolated contexts: high-volume operations (reading 30 files, grepping AST, running test suites) execute in separate worktrees, so 400k+ tokens stay in subagents while the primary session keeps a pristine 14k context
Plan on high. Delegate on medium. Keep Opus on call.
- the full subagent swarm setup
> Opus 5.5 on call in ~/.claude/agents/code-reviewer.md catches deep architectural traps
> Sonnet 5.5 lead drives edits, writes core logic, and executes test harnesses
> Haiku 5.5 subagents swarm AST parsing, grep scans, and API docs in parallel at 340 tok/s
> isolation: worktree guarantees test runs and scratch edits never pollute your git tree
> isolated context windows shield the primary session from 400k+ tokens of verbose output
paste the setup and this prompt into Claude Code below:
"Configure my Claude Code workspace for hierarchical subagent orchestration:
1. Audit ~/.claude/settings.json and project config for model roles fitting lead, subagents, and reviewer:
> Pin main session model: sonnet with effortLevel: h
RT @zeeg: if it is helpful for anyone who is curious if their luna use case might be doable on haiku, id assume the default is prob not (th…RT @zeeg: if it is helpful for anyone who is curious if their luna use case might be doable on haiku, id assume the default is prob not (th…
i can't comprehend why people run claude code on default settings
one settings change and every model gets its own job:
→ opus plans every task and advises when stuck
→ sonnet 5.5 writes the actual code
→ haiku handles every file search
→ a critic agent tries to break your diff before you merge
Opus thinks. The chea…i can't comprehend why people run claude code on default settings
one settings change and every model gets its own job:
→ opus plans every task and advises when stuck
→ sonnet 5.5 writes the actual code
→ haiku handles every file search
→ a critic agent tries to break your diff before you merge
Opus thinks. The cheap models grind.
paste this into claude code 👇
Add this to ~/.claude/settings.json. Keep every hook, permission and plugin I already have. Then create the two agent files below and show me the diff.
```json
{
"model": "opusplan",
"advisorModel": "opus",
"effortLevel": "medium",
"modelSettings": {
"claude-opus-5-5": { "effortLevel": "high" },
"claude-sonnet-5-5": { "effortLevel": "medium" }
},
"env": {
"CLAUDE_CODE_SUBAGENT_MODEL": "haiku"
}
}
```
~/.claude/agents/scout.md
```
---
name: scout
description: finds code fast. returns file paths and line numbers only.
model: haiku
tools: Read, Grep, Glob
---
Locate exactly what was asked. Return file:line. No fixes. No opinions.
```
~/.claude/agents/critic.md
```
---
name: critic
description: reviews a diff or plan before any task is called done
model: opus
tools: Read, Grep, Bash
---
Assume this is broken. Find how it fails. One line per issue, worst first.
```
Claude Haiku 5.5 is currently FREE 😳
Limited time on TokenHarbor.
No API key needed to try → just open and chat.
If you hit the limit, switch account and keep going.
Link: https://t.co/CHRl1Wp6qn
This won’t last long. https://t.co/Nez0Z3CTzKClaude Haiku 5.5 is currently FREE 😳
Limited time on TokenHarbor.
No API key needed to try → just open and chat.
If you hit the limit, switch account and keep going.
Link: https://t.co/CHRl1Wp6qn
This won’t last long. https://t.co/Nez0Z3CTzK
Anthropic clearly has had some sort of research breakthrough the last month
Opus, Sonnet, and Haiku 5.5 are all best in their class
Plus the speed and most importantly the usage are both INCREDIBLE
The best way I've found of using Claude is in Herdr
It allows you to open multiple Claude Code agents in tabs/projects…Anthropic clearly has had some sort of research breakthrough the last month
Opus, Sonnet, and Haiku 5.5 are all best in their class
Plus the speed and most importantly the usage are both INCREDIBLE
The best way I've found of using Claude is in Herdr
It allows you to open multiple Claude Code agents in tabs/projects
The most I've been able to multi task with agents in my life
It's completely free and open source btw (and no I'm not sponsored, I just love their product)
Open up a few instances of Opus in Herdr. Tell it to use Haiku as subagents. You'll multitask like crazy and get WAY better results
Happy Friday night coding!
This is no longer true.
Haiku has placed Anthropic in a position of complete and utter domination of the market. https://t.co/62CSpaumTrThis is no longer true.
Haiku has placed Anthropic in a position of complete and utter domination of the market. https://t.co/62CSpaumTr
WHOEVER LEAKED THIS CLAUDE CODE SETUP DESERVES A RAISE
I thought the best way to use Claude Code was to just throw Opus at everything.
Apparently not.
Anthropic’s setup is much more interesting:
→ Sonnet 5.5 runs the main session and actually builds the thing
→ Haiku 5.5 swarms the repo for file discovery, AST scan…WHOEVER LEAKED THIS CLAUDE CODE SETUP DESERVES A RAISE
I thought the best way to use Claude Code was to just throw Opus at everything.
Apparently not.
Anthropic’s setup is much more interesting:
→ Sonnet 5.5 runs the main session and actually builds the thing
→ Haiku 5.5 swarms the repo for file discovery, AST scans, grep, and docs
→ Opus 5.5 stays in the background as the architect
→ Opus gets called in before big plans, after repeated test failures, and before shipping
→ JEV handles the boring mechanical routing in ~16ms so frontier models only wake up when real reasoning is needed
so you get:
Opus reviews. Sonnet builds. Haiku swarms. JEV routes.
the interesting part isn’t using multiple models.
it’s giving each model a different job.
Instead of burning Opus context on finding files and running routine commands, you save it for the moments where architectural judgment actually matters.
Plan on high.Delegate on medium. Keep Opus on call.
❤️ 𝐋𝐈𝐊𝐄 ❤️
🔁 𝐍𝐄𝐄𝐃 𝐑𝐓 🔁
💬 𝐂𝐎𝐌𝐌𝐄𝐍𝐓 💬
🔖 𝐃𝐎 𝐁𝐎𝐎𝐊𝐌𝐀𝐑𝐊 🔖
👤 𝐅𝐎𝐋𝐋𝐎𝐖 𝐌𝐄 👤 @fran_alex897