Model APIs - 42 entries
xAI API rate limits
xAI's official docs, read on November 6, 2026, set rate limits for the xAI API across Tier 0, Tier 1, Tier 2, Tier 3 and Tier 4. For grok-4.7, grok-4.6, grok-4.5, Tier 0 is listed at 150 RPS and 50M TPM, and Tier 4 at 500 RPS and 100M TPM.
Last verified from docs.x.ai. What the docs leave out
- Tiers
- 5
- Tables
- 10
- Entries
- 42
- Verified
- Nov 6, 2026
Limits by model, endpoint and plan
Values are copied exactly as the docs show them, units included, in the docs' own order. A dash means the docs list no value for that row.
All 10 tables on this page
- grok-4.7, grok-4.6, grok-4.5
- grok-4.3, grok-4.20-0309-reasoning, grok-4.20-0309-non-reasoning, grok-build-0.1
- grok-4.20-multi-agent-0309
- grok-imagine-image, grok-imagine-image-2.0
- grok-imagine-video, grok-imagine-video-1.5, grok-imagine-video-1.5-lite
- grok-voice-3.0
- Text to Speech
- Speech to Text
- All text, embedding and voice models (team-level tier)
- Batch API (batch-enabled models)
| Tier | RPS |
|---|---|
| Tier 0 | 150 |
| |
| Tier 1 | 172 |
| |
| Tier 2 | 208 |
| |
| Tier 3 | 312 |
| |
| Tier 4 | 500 |
| |
- Tier 0: Language models. RPS is derived from the per-minute request budget (RPM / 48, with a minimum of 2). All tokens count toward TPM: prompt (text, image, audio), completion, reasoning and cached prompt tokens. Exceeding any limit returns 429 Too Many Requests. Docs page last updated October 5, 2026; personalized team limits are shown on the Models page in the xAI Console.
- Tier 1, Tier 2, Tier 3 and Tier 4: Same caveats as the tier-0 record for this scope.
grok-4.3, grok-4.20-0309-reasoning, grok-4.20-0309-non-reasoning, grok-build-0.1
5 tiersModel pages: Grok 4.3, Grok Build 0.1
| Tier | RPS |
|---|---|
| Tier 0 | 37 |
| |
| Tier 1 | 50 |
| |
| Tier 2 | 75 |
| |
| Tier 3 | 125 |
| |
| Tier 4 | 208 |
| |
- Tier 0: Language models. RPS is derived from the per-minute request budget (RPM / 48, with a minimum of 2). All tokens count toward TPM: prompt (text, image, audio), completion, reasoning and cached prompt tokens. Exceeding any limit returns 429 Too Many Requests. Docs page last updated October 5, 2026; personalized team limits are shown on the Models page in the xAI Console.
- Tier 1, Tier 2, Tier 3 and Tier 4: Same caveats as the tier-0 record for this scope.
grok-4.20-multi-agent-0309
5 tiers
| Tier | RPS |
|---|---|
| Tier 0 | 9 |
| |
| Tier 1 | 12 |
| |
| Tier 2 | 18 |
| |
| Tier 3 | 31 |
| |
| Tier 4 | 56 |
| |
- Tier 0: Language model. RPS is derived from the per-minute request budget (RPM / 48, with a minimum of 2). All tokens count toward TPM. Exceeding any limit returns 429 Too Many Requests. Docs page last updated October 5, 2026; personalized team limits are shown on the Models page in the xAI Console.
- Tier 1, Tier 2, Tier 3 and Tier 4: Same caveats as the tier-0 record for this scope.
grok-imagine-image, grok-imagine-image-2.0
5 tiers
| Tier | RPS |
|---|---|
| Tier 0 | 6 |
| |
| Tier 1 | 12 |
| |
| Tier 2 | 25 |
| |
| Tier 3 | 50 |
| |
| Tier 4 | 100 |
| |
- Tier 0: Image generation models. No TPM limit is published (shown as a dash). Exceeding any limit returns 429 Too Many Requests. Docs page last updated October 5, 2026; personalized team limits are shown on the Models page in the xAI Console.
- Tier 1, Tier 2, Tier 3 and Tier 4: Same caveats as the tier-0 record for this scope.
grok-imagine-video, grok-imagine-video-1.5, grok-imagine-video-1.5-lite
5 tiers
| Tier | RPS |
|---|---|
| Tier 0 | 10 |
| |
| Tier 1 | 20 |
| |
| Tier 2 | 39 |
| |
| Tier 3 | 79 |
| |
| Tier 4 | 158 |
| |
- Tier 0: Video generation models. No TPM limit is published (shown as a dash). Exceeding any limit returns 429 Too Many Requests. Docs page last updated October 5, 2026; personalized team limits are shown on the Models page in the xAI Console.
- Tier 1, Tier 2, Tier 3 and Tier 4: Same caveats as the tier-0 record for this scope.
grok-voice-3.0
5 tiers
| Tier | RPS |
|---|---|
| Tier 0 | not published (dash in docs table) |
| |
| Tier 1 | not published (dash in docs table) |
| |
| Tier 2 | not published (dash in docs table) |
| |
| Tier 3 | not published (dash in docs table) |
| |
| Tier 4 | not published (dash in docs table) |
| |
- Tier 0: Voice & Audio endpoints are limited by RPS and concurrent sessions (CST) rather than tokens; Speech to Speech is limited by concurrent sessions alone. Docs page last updated October 5, 2026; personalized team limits are shown on the Models page in the xAI Console.
- Tier 1, Tier 2, Tier 3 and Tier 4: Same caveats as the tier-0 record for this scope.
Text to Speech
5 tiers
| Tier | RPS |
|---|---|
| Tier 0 | 50 |
| |
| Tier 1 | 50 |
| |
| Tier 2 | 100 |
| |
| Tier 3 | 250 |
| |
| Tier 4 | 500 |
| |
- Tier 0: Voice & Audio endpoint, limited by requests per second and concurrent sessions (CST) rather than tokens. Docs page last updated October 5, 2026; personalized team limits are shown on the Models page in the xAI Console.
- Tier 1, Tier 2, Tier 3 and Tier 4: Same caveats as the tier-0 record for this scope.
Speech to Text
5 tiers
| Tier | RPS |
|---|---|
| Tier 0 | 10 |
| |
| Tier 1 | 10 |
| |
| Tier 2 | 20 |
| |
| Tier 3 | 30 |
| |
| Tier 4 | 40 |
| |
- Tier 0: Voice & Audio endpoint, limited by requests per second and concurrent sessions (CST) rather than tokens. Docs page last updated October 5, 2026; personalized team limits are shown on the Models page in the xAI Console.
- Tier 1, Tier 2, Tier 3 and Tier 4: Same caveats as the tier-0 record for this scope.
All text, embedding and voice models (team-level tier)
5 tiers - Tier 0 to Tier 4, Enterprise
| Tier | spend threshold |
|---|---|
| Tier 0 | $0 (default) |
| Tier 1 | $50 |
| Tier 2 | $250 |
| Tier 3 | $1,000 |
| Tier 4 | $5,000 |
| Metric | Limit |
|---|---|
| Enterprise | Available on request |
| RPS derivation | RPM / 48, with a minimum of 2 |
- Tier is based on cumulative spend on the xAI API since January 1, 2026. Qualification is based on total revenue received through prepaid credit purchases or successfully fulfilled invoices. Once you qualify for a tier, you stay there permanently; tiers never downgrade. Limits scale exponentially with tier. Over-limit responses: HTTP 429 (REST) or grpc.StatusCode.RESOURCE_EXHAUSTED (xai_sdk); docs recommend exponential backoff.
Batch API (batch-enabled models)
| Metric | Limit |
|---|---|
| Batch creation rate | 2 batch creations per second per team |
| Add-batch-requests API calls | 1000 every 30 seconds (rolling, shared across all batches in the team) |
| Batches per team | unlimited |
| Requests per batch | unlimited in theory; >100,000 requests may be throttled |
| Max payload per request | 25MB |
| JSONL file max size | 200 MB with up to 50,000 requests |
| Media result URL expiry | 1 hour |
- Batch requests don't count towards rate limits. Most batch requests complete within 24 hours (best effort, not guaranteed). File-based batches are sealed after creation. Non-batch-enabled models are rejected with "not supported for batch processing".
How to raise your limits
grok-4.7, grok-4.6, grok-4.5, grok-4.3, grok-4.20-0309-reasoning, grok-4.20-0309-non-reasoning, grok-build-0.1 and 4 more
Automatic tier upgrade on cumulative spend; request an increase in the xAI Console (Models page); enterprise capacity via sales@x.ai. See tier-thresholds record.
grok-imagine-image, grok-imagine-image-2.0 and grok-imagine-video, grok-imagine-video-1.5, grok-imagine-video-1.5-lite
Per the docs: 'Rate limit tiers apply to text, embedding, and voice models. For increases to Imagine API limits, contact sales@x.ai.'
All text, embedding and voice models (team-level tier)
Tiers upgrade automatically based on cumulative spend on the xAI API since January 1, 2026 (Tier 1 $50, Tier 2 $250, Tier 3 $1,000, Tier 4 $5,000; Enterprise available on request); tiers never downgrade. Request an increase through the xAI Console (Models page) for higher limits without additional spend or limits beyond Tier 4; for enterprise-grade capacity email sales@x.ai.
What the docs leave out
- Per-model RPM values are not published on the rate-limits page. It gives per-model limits on two dimensions, RPS and TPM, and says RPS is derived from the per-minute request budget (RPM / 48, with a minimum of 2).
- Embedding models: the docs say rate limit tiers apply to embedding models, but the public per-model table lists no embedding model.
- Enterprise tier limits are not published ("Available on request"). Each team's current tier and per-model limits can be viewed on the Models page of the xAI Console (console.x.ai/team/default/models).
- Imagine (image/video) models show RPS per tier in the public table, yet the docs say rate limit tiers apply to text, embedding and voice models and that Imagine increases go through sales@x.ai. This conflict is stated, not resolved, in the source.
- 2026-11-06: the Image Generation table now lists grok-imagine-image and grok-imagine-image-2.0. grok-imagine-image-quality was removed from the image rows because the page no longer lists it.
FAQ
What are the xAI API rate limits?
xAI's official docs, read on November 6, 2026, set rate limits for the xAI API across Tier 0, Tier 1, Tier 2, Tier 3 and Tier 4. For grok-4.7, grok-4.6, grok-4.5, Tier 0 is listed at 150 RPS and 50M TPM, and Tier 4 at 500 RPS and 100M TPM.
How do I raise my xAI API rate limits?
Automatic tier upgrade on cumulative spend; request an increase in the xAI Console (Models page); enterprise capacity via sales@x.ai. See tier-thresholds record. Per the docs: 'Rate limit tiers apply to text, embedding, and voice models. For increases to Imagine API limits, contact sales@x.ai.' Tiers upgrade automatically based on cumulative spend on the xAI API since January 1, 2026 (Tier 1 $50, Tier 2 $250, Tier 3 $1,000, Tier 4 $5,000; Enterprise available on request); tiers never downgrade. Request an increase through the xAI Console (Models page) for higher limits without additional spend or limits beyond Tier 4; for enterprise-grade capacity email sales@x.ai.
Where do these xAI API numbers come from?
Every value is copied verbatim from xAI's official documentation at docs.x.ai, read on November 6, 2026. This page lists all 42 entries in 10 tables, each linked to the page it came from.
Related reading
- How to Cut Your LLM API Bill in 2026: Prompt Caching, Batch and Flex Tiers Compared Across 15 ProvidersCaching a shared prompt cut our worked example by about 78-80% on four models; batch takes 50% off at most providers but 40% at Moonshot and 20% at xAI. Only Anthropic, Fireworks and xAI say the two stack. The math, the assumptions and the quotes.
- Prompt Caching Across AI Providers: Prices, TTLs and SetupPrompt caching across OpenAI, Anthropic, Gemini, DeepSeek, Mistral, xAI and Qwen: prices, TTLs, setup code. A reused 100K prefix costs 71-93% less.
- Grok Bot Has the Right Shape: Four Primitives, Nothing ElseGrok Bot ships four primitives that compose - a text thread, its own cloud computer, a chief of staff over Bots, and show-it-once routines - and nothing else.
- Grok 4.5 in 10 Minutes: xAI's Fastest Model, 500K Context, and Build-Mode IntegrationGrok 4.5 in 10 minutes: xAI's flagship model with a 500K context window, function calling, structured outputs and build-mode agents, plus verified pricing.
- Text-to-Speech APIs for Developers in 2026: What to Actually UseText-to-speech APIs compared for developers: OpenAI, ElevenLabs, xAI Grok and Cartesia on quality, latency, price, streaming and voice cloning policies.
- AI API 429, 529 and 503 Errors: Retry Rules That Match the DocsWhich 429, 529 and 503 errors to retry across OpenAI, Claude, Gemini and Bedrock, how to read Retry-After, and a typed wrapper that skips quota errors.
Sources
Last verified . Every value is copied verbatim from the official pages below, read on that date.
Newsletter
Know when limits change
One email per week. New model releases, pricing and limit changes, and what they mean for what you build.