fix: alibaba provider default endpoint and model list by teknium1 · Pull Request #3484 · NousResearch/hermes-agent

teknium1 · 2026-03-28T04:55:47Z

Summary

Fixes the Alibaba/DashScope provider configuration which was using an Anthropic-format endpoint (dashscope-intl.aliyuncs.com/apps/anthropic) as the default base URL, but routing through the OpenAI SDK (chat_completions api_mode). The OpenAI SDK appends /chat/completions to the base URL, producing .../apps/anthropic/chat/completions which 404s.

Changes

auth.py — Default inference_base_url changed to https://coding-intl.dashscope.aliyuncs.com/v1 (OpenAI-compatible endpoint on the DashScope Coding platform).

models.py — Updated curated model list:

Removed models unavailable on coding-intl: qwen3-max, qwen-plus-latest, qwen3.5-flash, qwen-vl-max
Added third-party models available on the platform: glm-5, glm-4.7, kimi-k2.5, MiniMax-M2.5
Added comments documenting the two DashScope services and how to override for classic DashScope keys

main.py — Updated provider picker description from "Anthropic-compatible" to "Qwen + multi-provider".

config.py — Updated env var descriptions to reflect the new default endpoint.

tests — Updated runtime provider resolution tests to match new default URL. Tests now cover:

Default: coding-intl /v1 → chat_completions mode
Override: /apps/anthropic → auto-detects anthropic_messages mode

How routing works

The existing auto-detection in runtime_provider.py (line 408) handles both endpoints correctly:

URLs ending in /anthropic → anthropic_messages api_mode → Anthropic SDK
URLs ending in /v1 → chat_completions api_mode → OpenAI SDK

Users with classic DashScope keys can override DASHSCOPE_BASE_URL to:

https://dashscope-intl.aliyuncs.com/compatible-mode/v1 (OpenAI-compat)
https://dashscope-intl.aliyuncs.com/apps/anthropic (Anthropic-compat)

Verified

All 8 models tested with both chat completions and tool calling on the coding-intl /v1 endpoint:

✔ qwen3.5-plus       ✔ tools
✔ qwen3-coder-plus   ✔ tools
✔ qwen3-coder-next   ✔ tools
✔ glm-5              ✔ tools
✔ glm-4.7            ✔ tools
✔ kimi-k2.5          ✔ tools
✔ MiniMax-M2.5       ✔ tools

6530 tests passed (1 pre-existing flaky failure unrelated to this PR).

- Change default inference_base_url from dashscope-intl Anthropic-compat endpoint to coding-intl OpenAI-compat /v1 endpoint. The old Anthropic endpoint 404'd when used with the OpenAI SDK (which appends /chat/completions to a /apps/anthropic base URL). - Update curated model list: remove models unavailable on coding-intl (qwen3-max, qwen-plus-latest, qwen3.5-flash, qwen-vl-max), add third-party models available on the platform (glm-5, glm-4.7, kimi-k2.5, MiniMax-M2.5). - URL-based api_mode auto-detection still works: overriding DASHSCOPE_BASE_URL to an /apps/anthropic endpoint automatically switches to anthropic_messages mode. - Update provider description and env var descriptions to reflect the coding-intl multi-provider platform. - Update tests to match new default URL and test the anthropic override path instead.

… pages Fixes found by auditing docs against recent PRs/commits: Critical (misleading): - hooks.md: Remove stale 'planned — not yet wired' markers for 4 hooks that are now active (#3542). Add correct callback signatures. - security.md: Update tirith verdict behavior — block verdicts now go through approval flow instead of hard-blocking (#3428). Add pkill/killall self-termination guard and gateway-run backgrounding patterns (#3593). New feature docs: - configuration.md: Add tool_use_enforcement section with value table (auto/true/false/list) from #3551/#3528. - configuration.md: Expand auxiliary config with per-task timeouts (compression 120s, web_extract 30s, approval 30s) from #3597. - api-server.md: Add /v1/health alias, Security Headers section, CORS details (Max-Age, SSE headers, Idempotency-Key) from #3572/#3573/#3576/#3580/#3530. Stale/incomplete: - configuration.md: Fix Alibaba model name qwen-plus -> qwen3.5-plus (#3484). - environment-variables.md: Specify actual DashScope default URL. - cli-commands.md: Add alibaba to --provider list. - fallback-providers.md: Add Alibaba/DashScope to provider table. - email.md: Document noreply/automated sender filtering (#3606). - toolsets-reference.md: Add 4 missing platform toolsets — matrix, mattermost, dingtalk, api-server (#3583). - skills.md: List default GitHub taps including garrytan/gstack (#3605).

… pages (#3618) Fixes found by auditing docs against recent PRs/commits: Critical (misleading): - hooks.md: Remove stale 'planned — not yet wired' markers for 4 hooks that are now active (#3542). Add correct callback signatures. - security.md: Update tirith verdict behavior — block verdicts now go through approval flow instead of hard-blocking (#3428). Add pkill/killall self-termination guard and gateway-run backgrounding patterns (#3593). New feature docs: - configuration.md: Add tool_use_enforcement section with value table (auto/true/false/list) from #3551/#3528. - configuration.md: Expand auxiliary config with per-task timeouts (compression 120s, web_extract 30s, approval 30s) from #3597. - api-server.md: Add /v1/health alias, Security Headers section, CORS details (Max-Age, SSE headers, Idempotency-Key) from #3572/#3573/#3576/#3580/#3530. Stale/incomplete: - configuration.md: Fix Alibaba model name qwen-plus -> qwen3.5-plus (#3484). - environment-variables.md: Specify actual DashScope default URL. - cli-commands.md: Add alibaba to --provider list. - fallback-providers.md: Add Alibaba/DashScope to provider table. - email.md: Document noreply/automated sender filtering (#3606). - toolsets-reference.md: Add 4 missing platform toolsets — matrix, mattermost, dingtalk, api-server (#3583). - skills.md: List default GitHub taps including garrytan/gstack (#3605).

) - Change default inference_base_url from dashscope-intl Anthropic-compat endpoint to coding-intl OpenAI-compat /v1 endpoint. The old Anthropic endpoint 404'd when used with the OpenAI SDK (which appends /chat/completions to a /apps/anthropic base URL). - Update curated model list: remove models unavailable on coding-intl (qwen3-max, qwen-plus-latest, qwen3.5-flash, qwen-vl-max), add third-party models available on the platform (glm-5, glm-4.7, kimi-k2.5, MiniMax-M2.5). - URL-based api_mode auto-detection still works: overriding DASHSCOPE_BASE_URL to an /apps/anthropic endpoint automatically switches to anthropic_messages mode. - Update provider description and env var descriptions to reflect the coding-intl multi-provider platform. - Update tests to match new default URL and test the anthropic override path instead.

… pages (NousResearch#3618) Fixes found by auditing docs against recent PRs/commits: Critical (misleading): - hooks.md: Remove stale 'planned — not yet wired' markers for 4 hooks that are now active (NousResearch#3542). Add correct callback signatures. - security.md: Update tirith verdict behavior — block verdicts now go through approval flow instead of hard-blocking (NousResearch#3428). Add pkill/killall self-termination guard and gateway-run backgrounding patterns (NousResearch#3593). New feature docs: - configuration.md: Add tool_use_enforcement section with value table (auto/true/false/list) from NousResearch#3551/NousResearch#3528. - configuration.md: Expand auxiliary config with per-task timeouts (compression 120s, web_extract 30s, approval 30s) from NousResearch#3597. - api-server.md: Add /v1/health alias, Security Headers section, CORS details (Max-Age, SSE headers, Idempotency-Key) from NousResearch#3572/NousResearch#3573/NousResearch#3576/NousResearch#3580/NousResearch#3530. Stale/incomplete: - configuration.md: Fix Alibaba model name qwen-plus -> qwen3.5-plus (NousResearch#3484). - environment-variables.md: Specify actual DashScope default URL. - cli-commands.md: Add alibaba to --provider list. - fallback-providers.md: Add Alibaba/DashScope to provider table. - email.md: Document noreply/automated sender filtering (NousResearch#3606). - toolsets-reference.md: Add 4 missing platform toolsets — matrix, mattermost, dingtalk, api-server (NousResearch#3583). - skills.md: List default GitHub taps including garrytan/gstack (NousResearch#3605).