Skip to content

[Bug] Error from provider (Console Go): Upstream request failed — Kimi K3 #37815

Closed as not planned

Description

@nyaa666

Description

Kimi K3 model appears in the model selection list but throws "Error from provider (Console Go): Upstream request failed" when selected.

This error only happens with Kimi K3 — other models on Console Go work fine.

Image

Key details:

  • The model is listed and selectable, but fails on any request
  • OpenCode Go limits are within normal range (~40% weekly, ~60% monthly) — not a quota issue
  • Config files are clean: no custom top_p, temperature, or other parameter overrides in any config or agent files
  • No top_p:0 found anywhere (unlike Error from provider (Console Go): Upstream request failed #37231 which was caused by that)
  • Other Console Go models (e.g. deepseek-v4-flash) work without issues
  • Changing API key does not help

OpenCode version

1.18.3

OS

Windows

Steps to reproduce

  1. Open a new session
  2. Select Kimi K3 from the model picker
  3. Send any message
  4. Receive: "Error from provider (Console Go): Upstream request failed"
    [...]
  5. Switch to another model (e.g. deepseek-v4-flash) — works normally

Related

#37231, #37056, #36889

Activity

  1. github-actions commented on Jul 19, 2026

    @github-actions
    Contributor

    This issue might be a duplicate of existing issues. Please check:

    If your issue is already tracked in one of these, please add any additional context there rather than keeping a separate report.

  2. ctml91 commented on Jul 21, 2026

    @ctml91

    Any solutions for this yet?

  3. Caffeine19 commented on Jul 21, 2026

    @Caffeine19

    Same here in the VSCode Copilot.

    Image
          {
            "id": "kimi-k3",
            "name": "Kimi K3",
            "url": "https://opencode.ai/zen/go/v1/chat/completions",
            "toolCalling": true,
            "vision": true,
            "maxInputTokens": 1048576,
            "maxOutputTokens": 16384,
            "modelOptions": {
              "temperature": 0,
              "top_p": 0
            }
          },
    
  4. gschadow commented on Jul 22, 2026

    @gschadow

    I have this error for one of 3 Kimi K3 sessions, started in the middle of the work at 39% context. And reverting just the last turn resolved it. I wish I could see the details of the error message somewhere in opencode.

  5. iamvaleriofantozzi commented on Jul 27, 2026

    @iamvaleriofantozzi

    I can reproduce this with a newly reset OpenCode Go plan, so this does not appear to be exhausted account quota or a client temperature override.

    I tested the Go endpoint directly using the same API key and a minimal payload:

    • kimi-k3, temperature omitted: HTTP 429 — Provider rate limit exceeded
    • kimi-k3, temperature: 1: HTTP 429 — Provider rate limit exceeded
    • kimi-k2.7-code, temperature omitted: HTTP 200
    • mimo-v2.5, temperature omitted: HTTP 200

    The Kimi K3 requests did not include OpenCode-specific mcp or top-level system fields. Changing the Go API key also did not resolve it. This points to a stale or incorrect model-specific Kimi K3 rate-limit bucket/routing state on Console Go. The generic Upstream request failed message appears alongside or after the K3-specific 429 responses.

  6. waynezhao12 commented on Jul 29, 2026

    @waynezhao12

    It seems to be an issue with content review on the OpenCode proxy. I’ve found that this error occurs whenever a request mentions Chinese leaders or contains political language. On my end, this only happens with Kimi models provided through OpenCode Go.

  7. anon192784 commented on Aug 24, 2026

    @anon192784

    Going to commands (ctrl+p) and changing a model to Mimo V2.5 Free worked for me (at least for now)

  8. danielpetrica commented on Aug 28, 2026

    @danielpetrica

    i have this issue with deepseek models too but not politica mention at all

  9. stantaneous commented on Sep 12, 2026

    @stantaneous

    facing similar issue with muse spark 1.3 with opencode go subscription.

    Image
  10. CasperOng commented on Sep 13, 2026

    @CasperOng

    Same issue with Luna

  11. stantaneous commented on Sep 13, 2026

    @stantaneous

    For me it was an issue because of the long tool name in one of my MCPs. The Responses API used by Muse Spark enforces name ≤ 64 chars. One of my MCP tool name is exactly 65 char. So i disabled it and it worked.

  12. kokoy13 commented on Sep 15, 2026

    @kokoy13

    i have same issues with that. i using model Muse Spark 1.3. i think the error because of overlimit session opencode. because when i try to restart the session, it work again

  13. thdxr commented on Oct 11, 2026

    @thdxr
    Member

    Thanks for the report! OpenCode 2.0 is now out and development is focused on V2, so we're closing issues filed against OpenCode 1. Please upgrade (opencode upgrade --major, details: https://anoma.ly/notes/opencode-2-0), and if this is still relevant on 2.0, reopen this issue or open a new one with your 2.0 version.

    — from 𝕺𝖕𝖊𝖓𝕮𝖔𝖉𝖊

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions