[guillaumegay13]

Automated sync between the Manifest gateway and this catalog (modelparams-agent).

How entries are verified: the local gateway loads this branch's catalog; for every declared param and test value the agent configures it through the gateway's model-params API, sends a request through the proxy pinned to the model, and checks the gateway applied the value upstream and got a 200 from that model. Params that fail are removed from the entry before it is filed.

This PR is rebuilt from main on every run and merged automatically once CI is green.

Added (4)

Entry	Params verified	Dropped	Corrected	Cloned from
anthropic/claude-sonnet-5-5.yaml	max_tokens output_config.effort	temperature top_p top_k thinking.type thinking.budget_tokens thinking.display	—	anthropic/claude-sonnet-4-6.yaml
alibaba/glm-5.2-fast-preview.yaml	max_completion_tokens temperature top_p extra_body.top_k extra_body.enable_thinking extra_body.thinking_budget	—	—	alibaba/glm-5.2.yaml
anthropic/claude-haiku-5-5.yaml	max_tokens thinking.type	temperature top_p top_k thinking.budget_tokens	thinking.type [disabled]	anthropic/claude-haiku-4-5.yaml
anthropic/claude-opus-5-5.yaml	max_tokens thinking.type thinking.display output_config.effort	temperature top_p top_k thinking.budget_tokens	thinking.type [adaptive]	anthropic/claude-opus-4-6.yaml

Evidence

models/anthropic/claude-sonnet-5-5.yaml

• baseline: 200 from anthropic/claude-sonnet-5-5 (api_key)
• max_tokens ✅ 4096, 1
• thinking.display ✅ "summarized", "omitted"
• output_config.effort ✅ "low", "medium", "high", "max"
• ❌ dropped temperature: temperature=1 → HTTP 400: To turn thinking off on this model, send "thinking": {"type": "between_tools"} instead of {"type": "disabled"}. The model does not think before responding. The
• ❌ dropped top_p: top_p=1 → HTTP 400: To turn thinking off on this model, send "thinking": {"type": "between_tools"} instead of {"type": "disabled"}. The model does not think before responding. The
• ❌ dropped top_k: top_k=0 → HTTP 400: To turn thinking off on this model, send "thinking": {"type": "between_tools"} instead of {"type": "disabled"}. The model does not think before responding. The
• ❌ dropped thinking.type: thinking.type="disabled" → HTTP 400: To turn thinking off on this model, send "thinking": {"type": "between_tools"} instead of {"type": "disabled"}. The model does not think before responding. The
• ❌ dropped thinking.budget_tokens: thinking.budget_tokens=4096 → HTTP 400: "thinking.type.enabled" is not supported for this model. Use "thinking.type.adaptive" and "output_config.effort" to control thinking behavior.
• ❌ dropped thinking.display: depends on dropped thinking.type

models/alibaba/glm-5.2-fast-preview.yaml

• baseline: 200 from qwen/glm-5.2-fast-preview (api_key)
• max_completion_tokens ✅ 1
• temperature ✅ 0, 1.9
• top_p ✅ 0, 1
• extra_body.top_k ✅ 20, 0
• extra_body.enable_thinking ✅ true, false
• extra_body.thinking_budget ✅ 1

models/anthropic/claude-haiku-5-5.yaml

• thinking.type: values [disabled, enabled] → [disabled]; default disabled (unchanged)
• vendor: ""thinking.type.enabled" is not supported for this model. Use "thinking.type.adaptive" and "output_config.effort" to control thinking behavior."
• proof through the gateway (draft catalog with the corrected entry): disabled ✅
• other params re-verified on the corrected entry: max_tokens ✅
• baseline: 200 from anthropic/claude-haiku-5-5 (api_key)
• max_tokens ✅ 4096, 1
• thinking.type ✅ "disabled"
• ❌ dropped temperature: temperature=0 → HTTP 400: temperature is deprecated for this model.
• ❌ dropped top_p: top_p=1 → HTTP 400: top_p is deprecated for this model.
• ❌ dropped top_k: top_k=0 → HTTP 400: top_k is deprecated for this model.
• ❌ dropped thinking.budget_tokens: gateway refused {"thinking.budget_tokens":4096,"thinking.type":"enabled"}: PUT /model-params -> HTTP 400: {"message":"Invalid value for param "thinking.type"","error":"Bad Request","statusCode":400}

models/anthropic/claude-opus-5-5.yaml

• thinking.type: values [disabled, adaptive, enabled] → [adaptive]; default disabled → adaptive
• vendor: ""thinking.type.disabled" is not supported for this model. Use "thinking.type.adaptive" and "output_config.effort" to control thinking behavior."
• vendor: ""thinking.type.enabled" is not supported for this model. Use "thinking.type.adaptive" and "output_config.effort" to control thinking behavior."
• proof through the gateway (draft catalog with the corrected entry): adaptive ✅
• note: default disabled refused and the vendor names no default; set to adaptive (nearest remaining value in the declared order) — a maintainer may prefer the vendor's documented default
• other params re-verified on the corrected entry: max_tokens thinking.display output_config.effort ✅
• baseline: 200 from anthropic/claude-opus-5-5 (api_key)
• max_tokens ✅ 4096, 1
• thinking.type ✅ "adaptive"
• thinking.display ✅ "summarized", "omitted"
• output_config.effort ✅ "low", "medium", "high", "max"
• ❌ dropped temperature: temperature=1 → gateway did not apply it upstream (recorded undefined)
• ❌ dropped top_p: top_p=1 → HTTP 400: top_p is deprecated for this model.
• ❌ dropped top_k: top_k=0 → HTTP 400: top_k is deprecated for this model.
• ❌ dropped thinking.budget_tokens: gateway refused {"thinking.budget_tokens":4096,"thinking.type":"enabled"}: PUT /model-params -> HTTP 400: {"message":"Invalid value for param "thinking.type"","error":"Bad Request","statusCode":400}

Needs a human: params now refused (11)

Re-verification found these declared params refused. Not removed — the param guard blocks removals and dropping a param is a maintainer's call.

• models/anthropic/claude-opus-4-6.yaml top_p: top_p=1 → gateway did not apply it upstream (recorded undefined)
• models/anthropic/claude-opus-4-6.yaml thinking.type thinking.budget_tokens: thinking.type="enabled" → HTTP 400: max_tokens must be greater than thinking.budget_tokens. Please consult our documentation at https://platform.claude.com/docs/en/build-with-claude/extended-t (same error for each)
• models/anthropic/claude-sonnet-4-6.yaml top_p: top_p=1 → gateway did not apply it upstream (recorded undefined)
• models/anthropic/claude-sonnet-4-6.yaml thinking.type thinking.budget_tokens: thinking.type="enabled" → HTTP 400: max_tokens must be greater than thinking.budget_tokens. Please consult our documentation at https://platform.claude.com/docs/en/build-with-claude/extended-t (same error for each)
• models/mistral/magistral-medium-latest.yaml prompt_mode: prompt_mode="reasoning" → HTTP 400: Reasoning prompt mode is not enabled for this model
• models/mistral/magistral-medium-latest.yaml response_format.type: response_format.type="json_schema" → HTTP 400: Schema response format type requires a json_schema
• models/mistral/mistral-medium-latest.yaml response_format.type: response_format.type="json_schema" → HTTP 400: Schema response format type requires a json_schema
• models/moonshot/kimi-k2.7-code.yaml response_format.type: response_format.type="json_schema" → HTTP 400: Invalid request: missing required parameter: 'response_format.json_schema'
• models/nvidia/gpt-oss-20b.yaml seed: seed=18446744073709552000 → HTTP 400: 1 validation error: {'type': 'less_than_equal', 'loc': ('body', 'seed'), 'msg': 'Input should be less than or equal to 9223372036854775807', 'input': 1844674407

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
modelparams.dev	Ready	Preview	Oct 9, 2026 5:46am UTC