Brief
LlamaIndex 0.14.16 adds built-in rate limiters for LLM and embedding calls
The 2026-03-10 release notes for llama-index-core 0.14.16 list two new rate-limiting features alongside smaller fixes. The notes describe the features but not their defaults, configuration or provider coverage.
The release notes for llama-index-core 0.14.16, dated 2026-03-10, list "Add token-bucket rate limiter for LLM and embedding API calls" (#20712) and, as a separate feature entry, "feat(rate_limiter): add SlidingWindowRateLimiter for strict per-minute caps" (#20799).
That is the detail likely to matter in daily work: a per-minute request cap becomes something the library offers rather than something you wrap around it, which is the practical difference when a workflow fans out many calls against a metered key. The notes do not state defaults, how either limiter is configured, which providers they apply to, or whether they are enabled by default.
The same release also lists a Multimodal LLMReranker (#20743), respect for the Retry-After header in the OpenAI embeddings retry decorator (llama-index-embeddings-openai 0.5.2, #20813), and a security fix adding RestrictedUnpickler to SimpleObjectNodeMapping, tagged CWE-502 (#20857).
Our reading
Rate limiting moving into the core library changes where teams spend effort: instead of hand-rolling throttles and retry logic around a provider client, a fan-out workflow can be capped from inside the framework. Anyone running multi-call pipelines against metered LLM or embedding keys, or against keys with hard per-minute quotas, has a reason to look at this release. The release notes alone do n…
What to do or watch
Pin 0.14.16 in a test environment, run a known fan-out pipeline against a metered key with a strict per-minute cap, and compare throttle and retry errors against the same run without the cap; watch for documented defaults and provider coverage in the project's rate-limiter documentation, which this evidence does not include.
Source details and supporting facts
Each line is stated by the page named above it.
Stated by github.com
- llama-index-core 0.14.16 release notes are dated 2026-03-10.
- The release adds a token-bucket rate limiter for LLM and embedding API calls (#20712).
- The release adds a SlidingWindowRateLimiter for strict per-minute caps (#20799).
- The release adds a Multimodal LLMReranker (#20743).
- llama-index-embeddings-openai 0.5.2 respects the Retry-After header in the OpenAI retry decorator (#20813).
- The release adds RestrictedUnpickler to SimpleObjectNodeMapping as a security fix tagged CWE-502 (#20857).
Sources
- LlamaIndex releasesText stored 17 September 2026
How this story was checked. Written from the 1 page listed above, stored 17 September 2026; claims checked against that stored text on 17 September 2026.
What that means
- 6 of 6 reported statements were confirmed against the page that carries them; the rest were removed rather than published.
- Figures in the text were required to appear in the stored source text: yes. Identifiers: yes.
- The check reads stored text only: no claim rests on a fresh look that did not happen.
- Where the reporting was silent, the text says so instead of filling the gap.