@@ -48,10 +48,12 @@ Live tests are split into two layers so we can isolate failures:
4848- `pnpm test:live` (or `OPENCLAW_LIVE_TEST=1` if invoking Vitest directly)
4949- Set `OPENCLAW_LIVE_MODELS=modern` (or `all`, alias for modern) to actually run this suite; otherwise it skips to keep `pnpm test:live` focused on gateway smoke
5050- How to select models:
51-- `OPENCLAW_LIVE_MODELS=modern` to run the modern allowlist (Opus/Sonnet 4.6+, GPT-5.x + Codex, Gemini 3, GLM 4.7, MiniMax M2.7, Grok 4)
51+- `OPENCLAW_LIVE_MODELS=modern` to run the modern allowlist (Opus/Sonnet 4.6+, GPT-5.2 + Codex, Gemini 3, GLM 4.7, MiniMax M2.7, Grok 4)
5252- `OPENCLAW_LIVE_MODELS=all` is an alias for the modern allowlist
53-- or `OPENCLAW_LIVE_MODELS="openai/gpt-5.4,openai-codex/gpt-5.5,anthropic/claude-opus-4-6,..."` (comma allowlist)
53+- or `OPENCLAW_LIVE_MODELS="openai/gpt-5.2,openai-codex/gpt-5.2,anthropic/claude-opus-4-6,..."` (comma allowlist)
5454- Modern/all sweeps default to a curated high-signal cap; set `OPENCLAW_LIVE_MAX_MODELS=0` for an exhaustive modern sweep or a positive number for a smaller cap.
55+- Exhaustive sweeps use `OPENCLAW_LIVE_TEST_TIMEOUT_MS` for the whole direct-model test timeout. Default: 60 minutes.
56+- Set `OPENCLAW_LIVE_MODEL_CONCURRENCY=10` to run direct-model probes in parallel. Default: 1.
5557- How to select providers:
5658- `OPENCLAW_LIVE_PROVIDERS="google,google-antigravity,google-gemini-cli"` (comma allowlist)
5759- Where keys come from:
@@ -80,7 +82,7 @@ Live tests are split into two layers so we can isolate failures:
8082- How to enable:
8183- `pnpm test:live` (or `OPENCLAW_LIVE_TEST=1` if invoking Vitest directly)
8284- How to select models:
83-- Default: modern allowlist (Opus/Sonnet 4.6+, GPT-5.x + Codex, Gemini 3, GLM 4.7, MiniMax M2.7, Grok 4)
85+- Default: modern allowlist (Opus/Sonnet 4.6+, GPT-5.2 + Codex, Gemini 3, GLM 4.7, MiniMax M2.7, Grok 4)
8486- `OPENCLAW_LIVE_GATEWAY_MODELS=all` is an alias for the modern allowlist
8587- Or set `OPENCLAW_LIVE_GATEWAY_MODELS="provider/model"` (or comma list) to narrow
8688- Modern/all gateway sweeps default to a curated high-signal cap; set `OPENCLAW_LIVE_GATEWAY_MAX_MODELS=0` for an exhaustive modern sweep or a positive number for a smaller cap.
@@ -115,7 +117,7 @@ openclaw models list --json
115117- Default provider/model: `claude-cli/claude-sonnet-4-6`
116118- Command/args/image behavior come from the owning CLI backend plugin metadata.
117119- Overrides (optional):
118-- `OPENCLAW_LIVE_CLI_BACKEND_MODEL="codex-cli/gpt-5.5"`
120+- `OPENCLAW_LIVE_CLI_BACKEND_MODEL="codex-cli/gpt-5.2"`
119121- `OPENCLAW_LIVE_CLI_BACKEND_COMMAND="/full/path/to/codex"`
120122- `OPENCLAW_LIVE_CLI_BACKEND_ARGS='["exec","--json","--color","never","--sandbox","read-only","--skip-git-repo-check"]'`
121123- `OPENCLAW_LIVE_CLI_BACKEND_IMAGE_PROBE=1` to send a real image attachment (paths are injected into the prompt).
@@ -128,7 +130,7 @@ Example:
128130129131```bash
130132OPENCLAW_LIVE_CLI_BACKEND=1 \
131- OPENCLAW_LIVE_CLI_BACKEND_MODEL="codex-cli/gpt-5.5" \
133+ OPENCLAW_LIVE_CLI_BACKEND_MODEL="codex-cli/gpt-5.2" \
132134 pnpm test:live src/gateway/gateway-cli-backend.live.test.ts
133135```
134136@@ -178,8 +180,8 @@ Notes:
178180- `OPENCLAW_LIVE_ACP_BIND_AGENT=gemini`
179181- `OPENCLAW_LIVE_ACP_BIND_AGENTS=claude,codex,gemini`
180182- `OPENCLAW_LIVE_ACP_BIND_AGENT_COMMAND='npx -y @agentclientprotocol/claude-agent-acp@<version>'`
181-- `OPENCLAW_LIVE_ACP_BIND_CODEX_MODEL=gpt-5.5`
182-- `OPENCLAW_LIVE_ACP_BIND_PARENT_MODEL=openai/gpt-5.4`
183+- `OPENCLAW_LIVE_ACP_BIND_CODEX_MODEL=gpt-5.2`
184+- `OPENCLAW_LIVE_ACP_BIND_PARENT_MODEL=openai/gpt-5.2`
183185- Notes:
184186- This lane uses the gateway `chat.send` surface with admin-only synthetic originating-route fields so tests can attach message-channel context without pretending to deliver externally.
185187- When `OPENCLAW_LIVE_ACP_BIND_AGENT_COMMAND` is unset, the test uses the embedded `acpx` plugin's built-in agent registry for the selected ACP harness agent.
@@ -220,7 +222,7 @@ Docker notes:
220222`agent` method:
221223- load the bundled `codex` plugin
222224- select `OPENCLAW_AGENT_RUNTIME=codex`
223-- send a first gateway agent turn to `openai/gpt-5.4` with the Codex harness forced
225+- send a first gateway agent turn to `openai/gpt-5.2` with the Codex harness forced
224226- send a second turn to the same OpenClaw session and verify the app-server
225227 thread can resume
226228- run `/codex status` and `/codex models` through the same gateway command
@@ -230,7 +232,7 @@ Docker notes:
230232 denied so the agent asks back
231233- Test: `src/gateway/gateway-codex-harness.live.test.ts`
232234- Enable: `OPENCLAW_LIVE_CODEX_HARNESS=1`
233-- Default model: `openai/gpt-5.4`
235+- Default model: `openai/gpt-5.2`
234236- Optional image probe: `OPENCLAW_LIVE_CODEX_HARNESS_IMAGE_PROBE=1`
235237- Optional MCP/tool probe: `OPENCLAW_LIVE_CODEX_HARNESS_MCP_PROBE=1`
236238- Optional Guardian probe: `OPENCLAW_LIVE_CODEX_HARNESS_GUARDIAN_PROBE=1`
@@ -248,7 +250,7 @@ OPENCLAW_LIVE_CODEX_HARNESS=1 \
248250 OPENCLAW_LIVE_CODEX_HARNESS_IMAGE_PROBE=1 \
249251 OPENCLAW_LIVE_CODEX_HARNESS_MCP_PROBE=1 \
250252 OPENCLAW_LIVE_CODEX_HARNESS_GUARDIAN_PROBE=1 \
251- OPENCLAW_LIVE_CODEX_HARNESS_MODEL=openai/gpt-5.4 \
253+ OPENCLAW_LIVE_CODEX_HARNESS_MODEL=openai/gpt-5.2 \
252254 pnpm test:live -- src/gateway/gateway-codex-harness.live.test.ts
253255```
254256@@ -279,13 +281,13 @@ Docker notes:
279281Narrow, explicit allowlists are fastest and least flaky:
280282281283- Single model, direct (no gateway):
282-- `OPENCLAW_LIVE_MODELS="openai/gpt-5.4" pnpm test:live src/agents/models.profiles.live.test.ts`
284+- `OPENCLAW_LIVE_MODELS="openai/gpt-5.2" pnpm test:live src/agents/models.profiles.live.test.ts`
283285284286- Single model, gateway smoke:
285-- `OPENCLAW_LIVE_GATEWAY_MODELS="openai/gpt-5.4" pnpm test:live src/gateway/gateway-models.profiles.live.test.ts`
287+- `OPENCLAW_LIVE_GATEWAY_MODELS="openai/gpt-5.2" pnpm test:live src/gateway/gateway-models.profiles.live.test.ts`
286288287289- Tool calling across several providers:
288-- `OPENCLAW_LIVE_GATEWAY_MODELS="openai/gpt-5.4,openai-codex/gpt-5.5,anthropic/claude-opus-4-6,google/gemini-3-flash-preview,zai/glm-4.7,minimax/MiniMax-M2.7" pnpm test:live src/gateway/gateway-models.profiles.live.test.ts`
290+- `OPENCLAW_LIVE_GATEWAY_MODELS="openai/gpt-5.2,openai-codex/gpt-5.2,anthropic/claude-opus-4-6,google/gemini-3-flash-preview,zai/glm-4.7,minimax/MiniMax-M2.7" pnpm test:live src/gateway/gateway-models.profiles.live.test.ts`
289291290292- Google focus (Gemini API key + Antigravity):
291293- Gemini (API key): `OPENCLAW_LIVE_GATEWAY_MODELS="google/gemini-3-flash-preview" pnpm test:live src/gateway/gateway-models.profiles.live.test.ts`
@@ -308,22 +310,22 @@ There is no fixed “CI model list” (live is opt-in), but these are the **reco
308310309311This is the “common models” run we expect to keep working:
310312311-- OpenAI (non-Codex): `openai/gpt-5.4` (optional: `openai/gpt-5.4-mini`)
312-- OpenAI Codex OAuth: `openai-codex/gpt-5.5`
313+- OpenAI (non-Codex): `openai/gpt-5.2`
314+- OpenAI Codex OAuth: `openai-codex/gpt-5.2`
313315- Anthropic: `anthropic/claude-opus-4-6` (or `anthropic/claude-sonnet-4-6`)
314316- Google (Gemini API): `google/gemini-3.1-pro-preview` and `google/gemini-3-flash-preview` (avoid older Gemini 2.x models)
315317- Google (Antigravity): `google-antigravity/claude-opus-4-6-thinking` and `google-antigravity/gemini-3-flash`
316318- Z.AI (GLM): `zai/glm-4.7`
317319- MiniMax: `minimax/MiniMax-M2.7`
318320319321Run gateway smoke with tools + image:
320-`OPENCLAW_LIVE_GATEWAY_MODELS="openai/gpt-5.4,openai-codex/gpt-5.5,anthropic/claude-opus-4-6,google/gemini-3.1-pro-preview,google/gemini-3-flash-preview,google-antigravity/claude-opus-4-6-thinking,google-antigravity/gemini-3-flash,zai/glm-4.7,minimax/MiniMax-M2.7" pnpm test:live src/gateway/gateway-models.profiles.live.test.ts`
322+`OPENCLAW_LIVE_GATEWAY_MODELS="openai/gpt-5.2,openai-codex/gpt-5.2,anthropic/claude-opus-4-6,google/gemini-3.1-pro-preview,google/gemini-3-flash-preview,google-antigravity/claude-opus-4-6-thinking,google-antigravity/gemini-3-flash,zai/glm-4.7,minimax/MiniMax-M2.7" pnpm test:live src/gateway/gateway-models.profiles.live.test.ts`
321323322324### Baseline: tool calling (Read + optional Exec)
323325324326Pick at least one per provider family:
325327326-- OpenAI: `openai/gpt-5.4` (or `openai/gpt-5.4-mini`)
328+- OpenAI: `openai/gpt-5.2`
327329- Anthropic: `anthropic/claude-opus-4-6` (or `anthropic/claude-sonnet-4-6`)
328330- Google: `google/gemini-3-flash-preview` (or `google/gemini-3.1-pro-preview`)
329331- Z.AI (GLM): `zai/glm-4.7`
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。