This introduces a new field `thinking_allowed` on `LanguageModelRequest`
which lets us control whether thinking should be enabled if the model
supports it.
We permit thinking in the Inline Assistant, Edit File tool and the Git
Commit message generator, this should make generation faster when using
a thinking model, e.g. `claude-sonnet-4-thinking`
Release Notes:
- N/A
* Updates to `zed_llm_client-0.8.5` which adds support for `retry_after`
when anthropic provides it.
* Distinguishes upstream provider errors and rate limits from errors
that originate from zed's servers
* Moves `LanguageModelCompletionError::BadInputJson` to
`LanguageModelCompletionEvent::ToolUseJsonParseError`. While arguably
this is an error case, the logic in thread is cleaner with this move.
There is also precedent for inclusion of errors in the event type -
`CompletionRequestStatus::Failed` is how cloud errors arrive.
* Updates `PROVIDER_ID` / `PROVIDER_NAME` constants to use proper types
instead of `&str`, since they can be constructed in a const fashion.
* Removes use of `CLIENT_SUPPORTS_EXA_WEB_SEARCH_PROVIDER_HEADER_NAME`
as the server no longer reads this header and just defaults to that
behavior.
Release notes for this is covered by #33275
Release Notes:
- N/A
---------
Co-authored-by: Richard Feldman <oss@rtfeldman.com>
Co-authored-by: Richard <richard@zed.dev>
This cleans up our settings to not include any `version` fields, as we
have an actual settings migrator now.
This PR removes `language_models > anthropic > version`,
`language_models > openai > version` and `agent > version`.
We had migration paths in the code for a long time, so in practice
almost everyone should be using the latest version of these settings.
Release Notes:
- Remove `version` fields in settings for `agent`, `language_models >
anthropic`, `language_models > openai`. Your settings will automatically
be migrated. If you're running into issues with this open an issue
[here](https://github.com/zed-industries/zed/issues)
This PR is in preparation for doing automatic retries for certain
errors, e.g. Overloaded. It doesn't change behavior yet (aside from some
granularity of error messages shown to the user), but rather mostly
changes some error handling to be exhaustive enum matches instead of
`anyhow` downcasts, and leaves some comments for where the behavior
change will be in a future PR.
Release Notes:
- N/A
This format is enabled for Google models as they seem to prefer it.
A relevant unit eval's pass rate has increased from 0.77 to 0.98.
Diff-fenced format looks like this (markdown fences and a line hint are
optional):
```diff
<<<<<<< SEARCH line=42
...
=======
...
>>>>>>> REPLACE
```
Release Notes:
- Agent: Gemini models now use the diff-fenced format when making edits
Some MCP servers expose tools that take absolute paths as arguments. To
interact with these, the agent needs to know the absolute path to the
project directories, not just their names. This PR changes the system
prompt to include the full path to each worktree, and updates some tool
descriptions to reflect this.
Todo:
* [x] Run evals, make sure assistant still understand how to specify
paths for tools, now that we include abs paths in the system prompt.
Release Notes:
- Improved the agent's ability to use MPC tools that require absolute
paths to files and directories in the project.
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
These changes help the agent edit files when `<old_text>` matches more
than one location.
First, the agent can specify an optional `<old_text line=XX>` parameter.
When this is provided and multiple matches exist, we use this hint to
identify the best match.
Second, when there is ambiguity in matches, we now return the agent a
more helpful message listing the line numbers of all possible matches.
Together, these changes should reduce the number of misplaced edits and
agent confusion.
I have ensured the LLM Worker works with these prompt changes.
Release Notes:
- Agent: Improved locating edits
Adds some jitter to avoid the issue that all requests will retry at
roughly the same time in eval where we have a lot of concurrent
requests.
Release Notes:
- N/A
Bubbles up rate limit information so that we can retry after a certain
duration if needed higher up in the stack.
Also caps the number of concurrent evals running at once to also help.
Release Notes:
- N/A
Replace hardcoded 0.10 threshold with configurable parameter and set
0.05 default for most tests, with 0.2 for from_pixels_constructor
eval that produces more mismatched tags.
Release Notes:
- N/A
We run the unit evals once a day in the middle of the night, and trigger
a Slack post if it fails.
Release Notes:
- N/A
---------
Co-authored-by: Oleksiy Syvokon <oleksiy.syvokon@gmail.com>
This change instructs models to wrap new file content in Markdown fences
and introduces a parser for this format. The reasons are:
1. This is the format we put a lot of effort into explaining in the
system prompt.
2. Gemini really prefers to do it.
3. It adds an option for a model to think before writing the content
The `eval_zode` pass rate for GEmini models goes from 0% to 100%. Other
models were already at 100%, this hasn't changed.
Release Notes:
- N/A
This change improves `eval_extract_handle_command_output` results for
all models:
Model | Pass rate before | Pass rate after
----------------------------|------------------|----------------
claude-3.7-sonnet | 0.96 | 0.98
gemini-2.5-pro | 0.35 | 0.86
gpt-4.1 | 0.81 | 1.00
Part of this improvement comes from more robust evaluation, which now
accepts multiple possible outcomes. Another part is from the prompt
adaptation: addressing common Gemini failure modes, adding a few-shot
example, and, in the final commit, auto-rewriting instructions for
clarity and conciseness.
This change still needs validation from larger end-to-end evals.
Release Notes:
- N/A
1. Add system prompt: this is how it's called from threads. Previously,
we were sending
2. Fix an issue with writing agent thought into a newly created empty
file.
Release Notes:
- N/A
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Co-authored-by: Antonio Scandurra <me@as-cii.com>
https://github.com/zed-industries/zed/issues/30972 brought up another
case where our context is not enough to track the actual source of the
issue: we get a general top-level error without inner error.
The reason for this was `.ok_or_else(|| anyhow!("failed to read HEAD
SHA"))?; ` on the top level.
The PR finally reworks the way we use anyhow to reduce such issues (or
at least make it simpler to bubble them up later in a fix).
On top of that, uses a few more anyhow methods for better readability.
* `.ok_or_else(|| anyhow!("..."))`, `map_err` and other similar error
conversion/option reporting cases are replaced with `context` and
`with_context` calls
* in addition to that, various `anyhow!("failed to do ...")` are
stripped with `.context("Doing ...")` messages instead to remove the
parasitic `failed to` text
* `anyhow::ensure!` is used instead of `if ... { return Err(...); }`
calls
* `anyhow::bail!` is used instead of `return Err(anyhow!(...));`
Release Notes:
- N/A
This eval checks that Edit Agent can create an empty file without
writing its thoughts into it. This issue is not specific to empty files,
but it's easier to reproduce with them.
For some mysterious reason, I could easily reproduce this issue roughly
90% of the time in actual Zed. However, once I extract the exact LLM
request before the failure point and generate from that, the
reproduction rate drops to 2%!
Things I've tried to make sure it's not a fluke: disabling prompt
caching, capturing the LLM request via a proxy server, running the
prompt on Claude separately from evals. Every time it was mostly giving
good outcomes, which doesn't match my actual experience in Zed.
At some point I discovered that simply adding one insignificant space or
a newline to the prompt suddenly results in an outcome I tried to
reproduce almost perfectly.
This weirdness happens even outside the Zed code base and even when
using a different subscription. The result is the same: an extra newline
or space changes the model behavior significantly enough, so that the
pass rate drops from 99% to 0-3%
I have no explanation to this.
Release Notes:
- N/A
1. The `edit_file` tool tended to use `create_or_overwrite` a bit too
often, leading to corruption of long files. This change replaces the
boolean flag with an `EditFileMode` enum, which helps Agent make a more
deliberate choice when overwriting files.
With this change, the pass rate of the new eval increased from 10% to
100%.
2. eval: Added ability to run eval on top of an existing thread. Threads
can now be loaded from JSON files in the `SerializedThread` format,
which makes it easy to use real threads as starting points for
tests/evals.
3. Don't try to restore tool cards when running in headless or eval mode
-- we don't have a window to properly do this.
Release Notes:
- N/A
This is very basic support for them. There are a number of other TODOs
before this is really a first-class supported feature, so not adding any
release notes for it; for now, this PR just makes it so that if
read_file tries to read a PNG (which has come up in practice), it at
least correctly sends it to Anthropic instead of messing up.
This also lays the groundwork for future PRs for more first-class
support for images in tool calls across more image file formats and LLM
providers.
Release Notes:
- N/A
---------
Co-authored-by: Agus Zubiaga <hi@aguz.me>
Co-authored-by: Agus Zubiaga <agus@zed.dev>
This allows us to debug the raw edits that were generated when people
report feedback, when running evals and when opening the thread as
Markdown.
Release Notes:
- Improved debug output for agent threads.
Release Notes:
- Fixed a bug that would prevent the agent from working over SSH.
---------
Co-authored-by: Nathan Sobo <nathan@zed.dev>
Co-authored-by: Richard Feldman <oss@rtfeldman.com>
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
Co-authored-by: Cole Miller <m@cole-miller.net>
This improves the new eval scenario by ~80% (`0.29` vs `0.525`) without
decreasing performance in the other evals.
Release Notes:
- Improved the performance of the `edit_file` tool.
Release Notes:
- Fixed a bug that would cause rejecting a hunk from the agent to delete
the file if the agent had decided to rewrite that file from scratch.
Nathan here: I also tacked on a bunch of UI refinement.
Release Notes:
- Introduced the ability to follow the agent around as it reads and
edits files.
---------
Co-authored-by: Nathan Sobo <nathan@zed.dev>
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
Refs #29733
This pull request introduces a new field to the `StreamingEditFileTool`
that lets the model create or overwrite a file in a streaming way. When
one of the `assistant.stream_edits` setting / `agent-stream-edits`
feature flag is enabled, we are going to disable the `CreateFileTool` so
that the agent model can only use `StreamingEditFileTool` for file
creation.
Release Notes:
- N/A
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Co-authored-by: Oleksiy Syvokon <oleksiy.syvokon@gmail.com>
This pull request introduces a new tool for streaming edits. The
short-term goal is for this tool to replace the existing `EditFileTool`,
but we want to get this out the door as soon as possible so that we can
start testing it.
`StreamingEditFileTool` is mutually exclusive with `EditFileTool`. It
will be enabled by default for anyone who has the `agent-stream-edits`
feature flag, as well as people that set `assistant.stream_edits` to
`true` in their settings.
### Implementation
Streaming is achieved by requesting a completion while the `edit_file`
tool gets called. We invoke the model by taking the existing
conversation with the agent and appending a prompt specifically tailored
for editing. In that prompt, we ask the model to produce a stream of
`<old_text>`/`<new_text>` tags. As the model streams text in, we
incrementally parse it and start editing as soon as we can.
### Evals
Note that, as part of this pull request, I also defined some new evals
that I used to drive the behavior of the recursive LLM call. To run
them, use this command:
```bash
cargo test --package=assistant_tools --features eval -- eval_extract_handle_command_output
```
Or comment out the `#[cfg_attr(not(feature = "eval"), ignore)]` macro.
I recommend running them one at a time, because right now we don't
really have a way of orchestrating of all these evals. I think we should
invest into that effort once the new agent panel goes live.
Release Notes:
- N/A
---------
Co-authored-by: Nathan Sobo <nathan@zed.dev>
Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
Co-authored-by: Oleksiy Syvokon <oleksiy.syvokon@gmail.com>
