# No model replies because Codex quota is exhausted and Ollama context is too small **URL:** \ **Category:** Troubleshooting\ **Tags:** compaction, models, ollama, rate-limits, troubleshooting\ **Created:** [September 14, 2026, 10:30pm UTC](https://community.openclaw.ai/t/no-model-replies-because-codex-quota-is-exhausted-and-ollama-context-is-too-small/49 "2026-09-14T22:30:51Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![ArchivedMember](https://avatars.discourse-cdn.com/v4/letter/a/e56c9b/32.png) [@ArchivedMember](https://community.openclaw.ai/u/ArchivedMember)\ **Post date:** [September 14, 2026, 10:30pm UTC](https://community.openclaw.ai/t/no-model-replies-because-codex-quota-is-exhausted-and-ollama-context-is-too-small/49/1 "2026-09-14T22:30:51Z")
> **Adapted community discussion:** Paraphrased and condensed from an anonymized public OpenClaw Discord support conversation from August 2026. Technical guidance was reviewed for OpenClaw 2026.9.3. My UI only says the agent run failed before replying. Restarting and starting a new conversation do not help, and it feels like every configured model is broken. ---
**Author:** ![OpenClawSupport](https://avatars.discourse-cdn.com/v4/letter/o/49beb7/32.png) [@OpenClawSupport](https://community.openclaw.ai/u/OpenClawSupport)\ **Post date:** [September 14, 2026, 10:30pm UTC](https://community.openclaw.ai/t/no-model-replies-because-codex-quota-is-exhausted-and-ollama-context-is-too-small/49/2 "2026-09-14T22:30:51Z")
The UI message is a wrapper. Run one small test while following the Gateway log, then inspect the model routes: ```bash openclaw --version openclaw status --all openclaw gateway status --deep openclaw models status openclaw logs --follow ``` Cancel repeated queued attempts and share only the sanitized error lines. ---
**Author:** ![OpenClawSupport](https://avatars.discourse-cdn.com/v4/letter/o/49beb7/32.png) [@OpenClawSupport](https://community.openclaw.ai/u/OpenClawSupport)\ **Post date:** [September 14, 2026, 10:30pm UTC](https://community.openclaw.ai/t/no-model-replies-because-codex-quota-is-exhausted-and-ollama-context-is-too-small/49/3 "2026-09-14T22:30:51Z")
The log shows two independent blockers. OpenAI/Codex is returning a usage-limit 429, while the local Ollama model cannot fit the current prompt in its effective context. Compaction is also pinned to the exhausted OpenAI route. Changing compaction reserve settings will not create more model context or restore provider quota. ---
**Author:** ![ArchivedMember](https://avatars.discourse-cdn.com/v4/letter/a/e56c9b/32.png) [@ArchivedMember](https://community.openclaw.ai/u/ArchivedMember)\ **Post date:** [September 14, 2026, 10:30pm UTC](https://community.openclaw.ai/t/no-model-replies-because-codex-quota-is-exhausted-and-ollama-context-is-too-small/49/4 "2026-09-14T22:30:51Z")
Model status confirms the local model is the primary, there are no fallbacks, and the OpenAI profile has no remaining usage in its current window. Compaction is hard-pinned to OpenAI. ---
**Author:** ![OpenClawSupport](https://avatars.discourse-cdn.com/v4/letter/o/49beb7/32.png) [@OpenClawSupport](https://community.openclaw.ai/u/OpenClawSupport)\ **Post date:** [September 14, 2026, 10:30pm UTC](https://community.openclaw.ai/t/no-model-replies-because-codex-quota-is-exhausted-and-ollama-context-is-too-small/49/5 "2026-09-14T22:30:51Z")
Remove the unavailable compaction override, choose a local model variant whose actual server context can fit OpenClaw’s prompt, and test it in a completely new conversation with a minimal tool profile. For Ollama, set `num_ctx` on a dedicated model variant only if the hardware can support it. Otherwise wait for quota reset or configure another authenticated provider. ---
**Author:** ![ArchivedMember](https://avatars.discourse-cdn.com/v4/letter/a/e56c9b/32.png) [@ArchivedMember](https://community.openclaw.ai/u/ArchivedMember)\ **Post date:** [September 14, 2026, 10:30pm UTC](https://community.openclaw.ai/t/no-model-replies-because-codex-quota-is-exhausted-and-ollama-context-is-too-small/49/6 "2026-09-14T22:30:51Z")
After restarting, the new larger-context Ollama variant appears as the default. The original smaller models are still listed too. ---
**Author:** ![OpenClawSupport](https://avatars.discourse-cdn.com/v4/letter/o/49beb7/32.png) [@OpenClawSupport](https://community.openclaw.ai/u/OpenClawSupport)\ **Post date:** [September 14, 2026, 10:30pm UTC](https://community.openclaw.ai/t/no-model-replies-because-codex-quota-is-exhausted-and-ollama-context-is-too-small/49/7 "2026-09-14T22:30:51Z")
That establishes that the larger variant is configured, while the smaller entries can remain. Refresh the UI, create a new thread, select the larger variant, and send a minimal test.