Abacus AI provider: tool calls fail with all models - models output commands instead of executing

I'm experiencing tool call failures with Abacus AI provider. The models output tool commands as text instead of executing them, regardless of which model I select.

Steps to reproduce:

Connect Abacus AI with valid API key
Launch OpenCode
Select any Abacus model
Send a request that requires tool use (e.g., asking to read a file)
Expected behavior: Tools execute and return results
Actual behavior: Models output tool calls as raw text (e.g., showing <tool_calls> blocks) instead of executing them
OpenCode version: 1.1.34
Operating System: fill in

Related issues:

Providers: Abacus connectivity error #6836 - Providers: Abacus connectivity error
abacus: some model names are incorrect & raw output in chat #8780 - abacus: some model names are incorrect & raw output in chat (closed but issue persists)

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

abacus: some model names are incorrect & raw output in chat #8780: abacus: some model names are incorrect & raw output in chat (already reported with the exact same problem - models outputting commands as raw text instead of executing)
Providers: Abacus connectivity error #6836: Providers: Abacus connectivity error (earlier Abacus provider connectivity issue)

Feel free to ignore if your specific case differs from these.

chore: Add supportsStructuredOutputs option to blablador config

Same issue here, this is an absolute pain. I got an Abacus Suscription as a fallback from other services but its useless :(

Abacus RouteLLM provider: Claude models return raw tool call XML instead of executing tools

Description
When using Abacus AI's RouteLLM API as a custom OpenAI-compatible provider, Claude models (e.g., claude-opus-4-5) return raw tool call markup instead of properly executing tools.

Environment
OpenCode version: latest
Provider: Custom @ai-sdk/openai-compatible
Endpoint: https://routellm.abacus.ai/v1/

Configuration
json
{
  "provider": {
    "abacus-routellm": {
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "https://routellm.abacus.ai/v1/",
        "apiKey": "..."
      },
      "models": {
        "claude-opus-4-5-20251101": { "name": "Claude Opus 4.5" }
      }
    }
  }
}

Expected Behavior
Tool calls should be parsed and executed normally.

Actual Behavior
Model returns raw XML/JSON in the response:

<tool_call_id>toolu_01SZTG42E49VsturJosSqzCt</tool_call_id>

I think this is a bug of Abacus RouteLLM, not OpenCode. I'm also having the same issue when using Abacus with CodeCompanion, some inspection of the HTTP calls shows that the actual tool calls are missing from choices delta when using streaming responses. Quick fix with CodeCompanion is to just disable streaming responses but I'm not sure if that is configurable in OpenCode?

Response from RouteLLM:

{
  "model": "gpt-5-nano",
  "created": 1769973939,
  "choices": [
    {
      "index": 0,
      "delta": {
        "content": "<tool_call_id>call_7XHvCMb6jfEdwuyU9BJpODh0</tool_call_id><function_call_name>apply_patch</function_call_name>",
        "role": "assistant"
      },
      "finish_reason": null,
      "logprobs": null
    }
  ]
}

Response from api.openai.com as reference:

{
  "id": "chatcmpl-D4X6TAkCSK9d5q1k8UVklhLyJVn9P",
  "object": "chat.completion.chunk",
  "created": 1769972933,
  "model": "gpt-5-nano-2025-08-07",
  "service_tier": "default",
  "system_fingerprint": null,
  "choices": [
    {
      "index": 0,
      "delta": {
        "role": "assistant",
        "content": null,
        "tool_calls": [
          {
            "index": 0,
            "id": "call_IxvQwArMy3Z0k77sftYvGPZD",
            "type": "function",
            "function": {
              "name": "apply_patch",
              "arguments": ""
            }
          }
        ],
        "refusal": null
      },
      "finish_reason": null
    }
  ],
  "obfuscation": "iYufaDra6htRX"
}

I have the same issue on openclaw.

We will address this on the Abacus.AI side in the next day or two.

@arvindsun

@arvindsun GLM has the same problem, Request to fix it together

closing since it isn't an opencode issue

I apologize if this is not an appropriate place, but I need to warn people somewhere. Please anyone reading this, do not use abucus for opencode. I asked it to create a tictactoe game and it burrned 11,757 credits in ONE go. It did not even do that much. This is not a sustainable API model. The exact same prompt with the exact same model on 3 others platforms was less than a cent.

We are fixing issues with cache keys and tool use soon

I apologize if this is not an appropriate place, but I need to warn people somewhere. Please anyone reading this, do not use abucus for opencode. I asked it to create a tictactoe game and it burrned 11,757 credits in ONE go. It did not even do that much. This is not a sustainable API model. The exact same prompt with the exact same model on 3 others platforms was less than a cent.

The XML outputting as text happens with Copilot Sonnet/Opus and ChatGPT 5.2 as well. This is a widespread issue and not model-specific.

Any updates? I cant get the tool to do any tool calls with Opus 4.5 or 4.6. Im using the IDE you guys developed

switched to provider zen (https://opencode.ai/zen) works smooth

Pronto solucionaremos problemas con las claves de caché y el uso de herramientas

Disculpad si este no es un lugar apropiado, pero necesito advertir a la gente en algún sitio. Por favor, cualquiera que lea esto, no use Abucus para opencode. Le pedí que creara un juego de Tictactoe y me arrecadó 11.757 créditos de UNA vez. Ni siquiera hizo mucho. Este no es un modelo de API sostenible. El mismo prompt exacto con el mismo modelo en otras 3 plataformas costaba menos de un céntimo.

alguna novedad de esto?

@arvindsun

Are there any updates on this? Would greatly appreciate a fix.

@arvindsun

We haven't seen issues with this lately - our api seems to work with openclaw and opencode - what is your setup?

Indeed it is working now! Sorry from this thread I assumed that this was still an issue. I can confirm that (MCP) tool calling is working now.

Thanks!

Navigation Menu

Repository navigation

Abacus AI provider: tool calls fail with all models - models output commands instead of executing #10412

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Abacus AI provider: tool calls fail with all models - models output commands instead of executing #10412

Description

Activity

github-actions commented on Jan 24, 2026

davidvv commented on Jan 29, 2026

studioarneboon commented on Feb 1, 2026

vekkt0r commented on Feb 1, 2026

TylonHH commented on Feb 1, 2026

arvindsun commented on Feb 3, 2026

LeoNardo-LB commented on Feb 3, 2026

rekram1-node commented on Feb 3, 2026

agluck91 commented on Feb 3, 2026

arvindsun commented on Feb 3, 2026

omerabr commented on Feb 5, 2026

lynchypin commented on Feb 11, 2026

studioarneboon commented on Feb 11, 2026

rober-tm commented on Feb 18, 2026

hd24-jspoerer commented on Mar 13, 2026

arvindsun commented on Mar 13, 2026

hd24-jspoerer commented on Mar 13, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions