@ai-sdk/gateway 4.0.90 → 4.0.92

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (10) hide show
  1. package/CHANGELOG.md +20 -0
  2. package/README.md +1 -1
  3. package/dist/index.d.ts +14 -8
  4. package/dist/index.js +1 -1
  5. package/docs/00-ai-gateway.mdx +29 -27
  6. package/package.json +2 -2
  7. package/src/gateway-image-model-settings.ts +1 -0
  8. package/src/gateway-language-model-settings.ts +3 -1
  9. package/src/gateway-provider-options.ts +17 -4
  10. package/src/gateway-speech-model-settings.ts +2 -0
package/CHANGELOG.md CHANGED
@@ -1,5 +1,25 @@
1
1
  # @ai-sdk/gateway
2
2
 
3
+ ## 4.0.92
4
+
5
+ ### Patch Changes
6
+
7
+ - 124b6ef: chore(provider/gateway): update gateway model settings files
8
+
9
+ ## 4.0.91
10
+
11
+ ### Patch Changes
12
+
13
+ - 2693319: Add Gemini 3.8 TTS support with structured speech metadata and per-turn speaker and style controls for prebuilt voices. Preserve native WAV responses without adding a second header, support explicit raw PCM, mu-law, and A-law output, and identify headerless audio formats correctly. Add the Gemini 3.8 speech model IDs to Google and Gateway types.
14
+
15
+ Share transcript and custom-voice inspection through the Google provider internal export, and reject empty speech transcripts before sending a request. Default newer and custom model IDs to structured speech while preserving the legacy format for Gemini 2.5 and 3.1.
16
+
17
+ - b73f2f9: feat(provider/gateway): add quantization conditions to the has provider option
18
+ - Updated dependencies [fe07867]
19
+ - Updated dependencies [a4b0940]
20
+ - Updated dependencies [771e74b]
21
+ - @ai-sdk/provider-utils@5.0.47
22
+
3
23
  ## 4.0.90
4
24
 
5
25
  ### Patch Changes
package/README.md CHANGED
@@ -33,7 +33,7 @@ import { gateway } from '@ai-sdk/gateway';
33
33
  import { generateText } from 'ai';
34
34
 
35
35
  const { text } = await generateText({
36
- model: gateway('xai/grok-4.7'),
36
+ model: gateway('spacexai/grok-4.7'),
37
37
  prompt:
38
38
  'Tell me about the history of the San Francisco Mission-style burrito.',
39
39
  });
package/dist/index.d.ts CHANGED
@@ -6,9 +6,9 @@ type GatewayEmbeddingModelId = 'alibaba/qwen3-embedding-0.6b' | 'alibaba/qwen3-e
6
6
 
7
7
  type GatewayEvaluationModelId = 'typesafe-ai/jev' | (string & {});
8
8
 
9
- type GatewayImageModelId = 'bfl/flux-2-flex' | 'bfl/flux-2-klein-4b' | 'bfl/flux-2-klein-9b' | 'bfl/flux-2-max' | 'bfl/flux-2-pro' | 'bfl/flux-kontext-max' | 'bfl/flux-kontext-pro' | 'bfl/flux-pro-1.0-fill' | 'bfl/flux-pro-1.1' | 'bfl/flux-pro-1.1-ultra' | 'bytedance/seedream-4.0' | 'bytedance/seedream-4.5' | 'bytedance/seedream-5.0-lite' | 'bytedance/seedream-5.0-pro' | 'meta/muse-image-1.0' | 'openai/gpt-image-1' | 'openai/gpt-image-1-mini' | 'openai/gpt-image-1.5' | 'openai/gpt-image-2' | 'openai/gpt-image-2.5-flare' | 'openai/gpt-image-2.5-sunburst' | 'prodia/flux-fast-schnell' | 'quiverai/arrow-1.1' | 'recraft/recraft-v2' | 'recraft/recraft-v3' | 'recraft/recraft-v4' | 'recraft/recraft-v4-pro' | 'recraft/recraft-v4.1' | 'recraft/recraft-v4.1-pro' | 'recraft/recraft-v4.1-utility' | 'recraft/recraft-v4.1-utility-pro' | 'spacexai/grok-imagine-image' | 'spacexai/grok-imagine-image-2.0' | (string & {});
9
+ type GatewayImageModelId = 'bfl/flux-2-flex' | 'bfl/flux-2-klein-4b' | 'bfl/flux-2-klein-9b' | 'bfl/flux-2-max' | 'bfl/flux-2-pro' | 'bfl/flux-kontext-max' | 'bfl/flux-kontext-pro' | 'bfl/flux-pro-1.0-fill' | 'bfl/flux-pro-1.1' | 'bfl/flux-pro-1.1-ultra' | 'bytedance/seedream-4.0' | 'bytedance/seedream-4.5' | 'bytedance/seedream-5.0-lite' | 'bytedance/seedream-5.0-pro' | 'meta/muse-image-1.0' | 'openai/gpt-image-1' | 'openai/gpt-image-1-mini' | 'openai/gpt-image-1.5' | 'openai/gpt-image-2' | 'openai/gpt-image-2.5-flare' | 'openai/gpt-image-2.5-sunburst' | 'prodia/flux-fast-schnell' | 'quiverai/arrow-1.1' | 'recraft/recraft-v2' | 'recraft/recraft-v3' | 'recraft/recraft-v4' | 'recraft/recraft-v4-pro' | 'recraft/recraft-v4.1' | 'recraft/recraft-v4.1-flash' | 'recraft/recraft-v4.1-pro' | 'recraft/recraft-v4.1-utility' | 'recraft/recraft-v4.1-utility-pro' | 'spacexai/grok-imagine-image' | 'spacexai/grok-imagine-image-2.0' | (string & {});
10
10
 
11
- type GatewayModelId = 'alibaba/qwen-3-14b' | 'alibaba/qwen-3-235b' | 'alibaba/qwen-3-30b' | 'alibaba/qwen-3-32b' | 'alibaba/qwen-3.6-max-preview' | 'alibaba/qwen3-235b-a22b-thinking' | 'alibaba/qwen3-coder' | 'alibaba/qwen3-coder-30b-a3b' | 'alibaba/qwen3-coder-next' | 'alibaba/qwen3-coder-plus' | 'alibaba/qwen3-max' | 'alibaba/qwen3-max-preview' | 'alibaba/qwen3-max-thinking' | 'alibaba/qwen3-next-80b-a3b-instruct' | 'alibaba/qwen3-next-80b-a3b-thinking' | 'alibaba/qwen3-vl-235b-a22b-instruct' | 'alibaba/qwen3-vl-instruct' | 'alibaba/qwen3-vl-thinking' | 'alibaba/qwen3.5-flash' | 'alibaba/qwen3.5-plus' | 'alibaba/qwen3.6-27b' | 'alibaba/qwen3.6-plus' | 'alibaba/qwen3.7-flash' | 'alibaba/qwen3.7-max' | 'alibaba/qwen3.7-plus' | 'alibaba/qwen3.8-2.4t-a95b' | 'alibaba/qwen3.8-27b' | 'alibaba/qwen3.8-flash' | 'alibaba/qwen3.8-max' | 'alibaba/qwen3.8-max-0902' | 'alibaba/qwen3.8-omni-flash' | 'amazon/nova-2-lite' | 'amazon/nova-lite' | 'amazon/nova-micro' | 'amazon/nova-pro' | 'anthropic/claude-3-haiku' | 'anthropic/claude-fable-5' | 'anthropic/claude-fable-5.1' | 'anthropic/claude-haiku-4.5' | 'anthropic/claude-opus-4' | 'anthropic/claude-opus-4.5' | 'anthropic/claude-opus-4.6' | 'anthropic/claude-opus-4.7' | 'anthropic/claude-opus-4.8' | 'anthropic/claude-opus-4.8-fast' | 'anthropic/claude-opus-5' | 'anthropic/claude-opus-5-fast' | 'anthropic/claude-opus-5.5' | 'anthropic/claude-sonnet-4' | 'anthropic/claude-sonnet-4.5' | 'anthropic/claude-sonnet-4.6' | 'anthropic/claude-sonnet-5' | 'arcee-ai/trinity-large-thinking' | 'bytedance/seed-1.6' | 'bytedance/seed-1.8' | 'bytedance/seed-2.1-turbo' | 'cohere/command-a' | 'deepseek/deepseek-r1' | 'deepseek/deepseek-v3.1' | 'deepseek/deepseek-v3.1-terminus' | 'deepseek/deepseek-v3.2' | 'deepseek/deepseek-v3.2-thinking' | 'deepseek/deepseek-v4-flash' | 'deepseek/deepseek-v4-flash-0731' | 'deepseek/deepseek-v4-flash-vision-exp' | 'deepseek/deepseek-v4-pro' | 'deepseek/deepseek-v4-pro-0813' | 'deepseek/deepseek-v4.1-flash' | 'google/gemini-2.5-flash' | 'google/gemini-2.5-flash-image' | 'google/gemini-2.5-flash-lite' | 'google/gemini-2.5-pro' | 'google/gemini-3-flash' | 'google/gemini-3-pro-image' | 'google/gemini-3.1-flash-image' | 'google/gemini-3.1-flash-image-preview' | 'google/gemini-3.1-flash-lite' | 'google/gemini-3.1-flash-lite-image' | 'google/gemini-3.1-pro-preview' | 'google/gemini-3.5-flash' | 'google/gemini-3.5-flash-lite' | 'google/gemini-3.6-flash' | 'google/gemini-3.7-flash' | 'google/gemini-3.8-flash' | 'google/gemini-omni-flash-preview' | 'google/gemma-4-26b-a4b-it' | 'google/gemma-4-31b-it' | 'inception/mercury-2' | 'inception/mercury-2.5' | 'inception/mercury-coder-small' | 'inclusionai/ling-3.0-flash' | 'inclusionai/ling-3.0-flash-fin' | 'inclusionai/ling-3.0-flash-fin-free' | 'inclusionai/ling-3.0-flash-sante' | 'inclusionai/ling-3.0-flash-sante-free' | 'inclusionai/ling-3.0-flash-vl' | 'inclusionai/ling-3.0-flash-vl-free' | 'inference-net/schematron-v2-small' | 'inference-net/schematron-v2-turbo' | 'interfaze/interfaze-beta' | 'meta/llama-3.1-70b' | 'meta/llama-3.1-8b' | 'meta/llama-3.3-70b' | 'meta/llama-4-maverick' | 'meta/llama-4-scout' | 'meta/muse-glimmer-30b' | 'meta/muse-spark-1.1' | 'meta/muse-spark-1.2' | 'meta/muse-spark-1.2-contributor' | 'meta/muse-spark-1.3' | 'meta/muse-spark-1.3-contributor' | 'minimax/minimax-m2' | 'minimax/minimax-m2.1' | 'minimax/minimax-m2.1-lightning' | 'minimax/minimax-m2.5' | 'minimax/minimax-m2.5-highspeed' | 'minimax/minimax-m2.7' | 'minimax/minimax-m2.7-highspeed' | 'minimax/minimax-m3' | 'mistral/codestral' | 'mistral/ministral-14b' | 'mistral/ministral-3b' | 'mistral/ministral-8b' | 'mistral/mistral-large-3' | 'mistral/mistral-medium-3.5' | 'mistral/mistral-nemo' | 'mistral/mistral-small' | 'mixedbread/toast-1' | 'moonshotai/kimi-k2' | 'moonshotai/kimi-k2-thinking' | 'moonshotai/kimi-k2.5' | 'moonshotai/kimi-k2.6' | 'moonshotai/kimi-k2.7-code' | 'moonshotai/kimi-k2.7-code-highspeed' | 'moonshotai/kimi-k3' | 'moonshotai/kimi-k3-fast' | 'morph/morph-v3-fast' | 'morph/morph-v3-large' | 'nvidia/nemotron-3-nano-30b-a3b' | 'nvidia/nemotron-3-super-120b-a12b' | 'nvidia/nemotron-3-ultra-550b-a55b' | 'nvidia/nemotron-3.5-lightning' | 'nvidia/nemotron-nano-12b-v2-vl' | 'nvidia/nemotron-nano-9b-v2' | 'openai/gpt-3.5-turbo' | 'openai/gpt-4-turbo' | 'openai/gpt-4.1' | 'openai/gpt-4.1-fast' | 'openai/gpt-4.1-mini' | 'openai/gpt-4.1-mini-fast' | 'openai/gpt-4.1-nano' | 'openai/gpt-4.1-nano-fast' | 'openai/gpt-4o' | 'openai/gpt-4o-fast' | 'openai/gpt-4o-mini' | 'openai/gpt-4o-mini-fast' | 'openai/gpt-5' | 'openai/gpt-5-codex' | 'openai/gpt-5-fast' | 'openai/gpt-5-mini' | 'openai/gpt-5-mini-fast' | 'openai/gpt-5-nano' | 'openai/gpt-5-pro' | 'openai/gpt-5.1-codex' | 'openai/gpt-5.1-codex-max' | 'openai/gpt-5.1-codex-mini' | 'openai/gpt-5.1-thinking' | 'openai/gpt-5.1-thinking-fast' | 'openai/gpt-5.2' | 'openai/gpt-5.2-codex' | 'openai/gpt-5.2-fast' | 'openai/gpt-5.2-pro' | 'openai/gpt-5.3-codex' | 'openai/gpt-5.3-codex-fast' | 'openai/gpt-5.4' | 'openai/gpt-5.4-fast' | 'openai/gpt-5.4-mini' | 'openai/gpt-5.4-mini-fast' | 'openai/gpt-5.4-nano' | 'openai/gpt-5.4-pro' | 'openai/gpt-5.5' | 'openai/gpt-5.5-fast' | 'openai/gpt-5.5-pro' | 'openai/gpt-5.6-luna' | 'openai/gpt-5.6-luna-fast' | 'openai/gpt-5.6-sol' | 'openai/gpt-5.6-sol-fast' | 'openai/gpt-5.6-terra' | 'openai/gpt-5.6-terra-fast' | 'openai/gpt-6-astra' | 'openai/gpt-6-astra-fast' | 'openai/gpt-6-luna' | 'openai/gpt-6-luna-fast' | 'openai/gpt-6-sol' | 'openai/gpt-6-sol-fast' | 'openai/gpt-oss-120b' | 'openai/gpt-oss-20b' | 'openai/gpt-oss-safeguard-120b' | 'openai/gpt-oss-safeguard-20b' | 'openai/o1' | 'openai/o3' | 'openai/o3-fast' | 'openai/o3-mini' | 'openai/o3-pro' | 'openai/o4-mini' | 'openai/o4-mini-fast' | 'perplexity/sonar' | 'perplexity/sonar-pro' | 'perplexity/sonar-reasoning-pro' | 'poolside/laguna-s-2.1' | 'poolside/laguna-s-2.1-free' | 'quiverai/arrow-2' | 'quiverai/arrow-2-telos' | 'sakana/fugu-max' | 'sakana/fugu-ultra' | 'sakana/fugu-ultra-v2' | 'sakana/namazu' | 'spacexai/grok-4.1-fast-non-reasoning' | 'spacexai/grok-4.1-fast-reasoning' | 'spacexai/grok-4.20-multi-agent' | 'spacexai/grok-4.20-multi-agent-beta' | 'spacexai/grok-4.20-non-reasoning' | 'spacexai/grok-4.20-non-reasoning-beta' | 'spacexai/grok-4.20-reasoning' | 'spacexai/grok-4.20-reasoning-beta' | 'spacexai/grok-4.3' | 'spacexai/grok-4.5' | 'spacexai/grok-4.6' | 'spacexai/grok-4.7' | 'spacexai/grok-build-0.1' | 'stepfun/step-3.5-flash' | 'stepfun/step-3.7-flash' | 'stepfun/step-5-preview' | 'tencent/hy-mt2-lite' | 'tencent/hy-mt2-plus' | 'tencent/hy-mt2-pro' | 'tencent/hy3' | 'tencent/hy4-preview' | 'thinkingmachines/inkling' | 'thinkingmachines/inkling-small' | 'xiaomi/mimo-v2.5' | 'xiaomi/mimo-v2.5-pro' | 'xiaomi/mimo-v2.6-flash' | 'xiaomi/mimo-v2.6-pro' | 'xiaomi/mimo-v2.6-pro-ultraspeed' | 'zai/glm-4.5' | 'zai/glm-4.5-air' | 'zai/glm-4.5v' | 'zai/glm-4.6' | 'zai/glm-4.7' | 'zai/glm-4.7-flash' | 'zai/glm-4.7-flashx' | 'zai/glm-5' | 'zai/glm-5-turbo' | 'zai/glm-5.1' | 'zai/glm-5.2' | 'zai/glm-5.2-fast' | 'zai/glm-5.3' | 'zai/glm-5.3-fast' | 'zai/glm-5.3-flash' | 'zai/glm-5.3-flashx' | 'zai/glm-5v-turbo' | (string & {});
11
+ type GatewayModelId = 'alibaba/qwen-3-14b' | 'alibaba/qwen-3-235b' | 'alibaba/qwen-3-30b' | 'alibaba/qwen-3-32b' | 'alibaba/qwen-3.6-max-preview' | 'alibaba/qwen3-235b-a22b-thinking' | 'alibaba/qwen3-coder' | 'alibaba/qwen3-coder-30b-a3b' | 'alibaba/qwen3-coder-next' | 'alibaba/qwen3-coder-plus' | 'alibaba/qwen3-max' | 'alibaba/qwen3-max-preview' | 'alibaba/qwen3-max-thinking' | 'alibaba/qwen3-next-80b-a3b-instruct' | 'alibaba/qwen3-next-80b-a3b-thinking' | 'alibaba/qwen3-vl-235b-a22b-instruct' | 'alibaba/qwen3-vl-instruct' | 'alibaba/qwen3-vl-thinking' | 'alibaba/qwen3.5-flash' | 'alibaba/qwen3.5-plus' | 'alibaba/qwen3.6-27b' | 'alibaba/qwen3.6-plus' | 'alibaba/qwen3.7-flash' | 'alibaba/qwen3.7-max' | 'alibaba/qwen3.7-plus' | 'alibaba/qwen3.8-2.4t-a95b' | 'alibaba/qwen3.8-27b' | 'alibaba/qwen3.8-flash' | 'alibaba/qwen3.8-max' | 'alibaba/qwen3.8-max-0902' | 'alibaba/qwen3.8-max-prime' | 'alibaba/qwen3.8-omni-flash' | 'amazon/nova-2-lite' | 'amazon/nova-lite' | 'amazon/nova-micro' | 'amazon/nova-pro' | 'anthropic/claude-3-haiku' | 'anthropic/claude-fable-5' | 'anthropic/claude-fable-5.1' | 'anthropic/claude-haiku-4.5' | 'anthropic/claude-opus-4' | 'anthropic/claude-opus-4.5' | 'anthropic/claude-opus-4.6' | 'anthropic/claude-opus-4.7' | 'anthropic/claude-opus-4.8' | 'anthropic/claude-opus-4.8-fast' | 'anthropic/claude-opus-5' | 'anthropic/claude-opus-5-fast' | 'anthropic/claude-opus-5.5' | 'anthropic/claude-opus-5.5-fast' | 'anthropic/claude-sonnet-4' | 'anthropic/claude-sonnet-4.5' | 'anthropic/claude-sonnet-4.6' | 'anthropic/claude-sonnet-5' | 'arcee-ai/trinity-large-thinking' | 'bytedance/seed-1.6' | 'bytedance/seed-1.8' | 'bytedance/seed-2.1-turbo' | 'cohere/command-a' | 'deepseek/deepseek-r1' | 'deepseek/deepseek-v3.1' | 'deepseek/deepseek-v3.1-terminus' | 'deepseek/deepseek-v3.2' | 'deepseek/deepseek-v3.2-thinking' | 'deepseek/deepseek-v4-flash' | 'deepseek/deepseek-v4-flash-0731' | 'deepseek/deepseek-v4-flash-vision-exp' | 'deepseek/deepseek-v4-pro' | 'deepseek/deepseek-v4-pro-0813' | 'deepseek/deepseek-v4.1-flash' | 'fireworks/ember-1' | 'google/gemini-2.5-flash' | 'google/gemini-2.5-flash-image' | 'google/gemini-2.5-flash-lite' | 'google/gemini-2.5-pro' | 'google/gemini-3-flash' | 'google/gemini-3-pro-image' | 'google/gemini-3.1-flash-image' | 'google/gemini-3.1-flash-image-preview' | 'google/gemini-3.1-flash-lite' | 'google/gemini-3.1-flash-lite-image' | 'google/gemini-3.1-pro-preview' | 'google/gemini-3.5-flash' | 'google/gemini-3.5-flash-lite' | 'google/gemini-3.6-flash' | 'google/gemini-3.7-flash' | 'google/gemini-3.8-flash' | 'google/gemini-omni-flash-preview' | 'google/gemma-4-26b-a4b-it' | 'google/gemma-4-31b-it' | 'inception/mercury-2' | 'inception/mercury-2.5' | 'inception/mercury-coder-small' | 'inclusionai/ling-3.0-flash' | 'inclusionai/ling-3.0-flash-fin' | 'inclusionai/ling-3.0-flash-fin-free' | 'inclusionai/ling-3.0-flash-sante' | 'inclusionai/ling-3.0-flash-sante-free' | 'inclusionai/ling-3.0-flash-vl' | 'inference-net/schematron-v2-small' | 'inference-net/schematron-v2-turbo' | 'interfaze/interfaze-beta' | 'meta/llama-3.1-70b' | 'meta/llama-3.1-8b' | 'meta/llama-3.3-70b' | 'meta/llama-4-maverick' | 'meta/llama-4-scout' | 'meta/muse-glimmer-30b' | 'meta/muse-spark-1.1' | 'meta/muse-spark-1.2' | 'meta/muse-spark-1.2-contributor' | 'meta/muse-spark-1.3' | 'meta/muse-spark-1.3-contributor' | 'minimax/minimax-m2' | 'minimax/minimax-m2.1' | 'minimax/minimax-m2.1-lightning' | 'minimax/minimax-m2.5' | 'minimax/minimax-m2.5-highspeed' | 'minimax/minimax-m2.7' | 'minimax/minimax-m2.7-highspeed' | 'minimax/minimax-m3' | 'mistral/codestral' | 'mistral/ministral-14b' | 'mistral/ministral-3b' | 'mistral/ministral-8b' | 'mistral/mistral-large-3' | 'mistral/mistral-medium-3.5' | 'mistral/mistral-nemo' | 'mistral/mistral-small' | 'mixedbread/toast-1' | 'moonshotai/kimi-k2' | 'moonshotai/kimi-k2-thinking' | 'moonshotai/kimi-k2.5' | 'moonshotai/kimi-k2.6' | 'moonshotai/kimi-k2.7-code' | 'moonshotai/kimi-k2.7-code-highspeed' | 'moonshotai/kimi-k3' | 'moonshotai/kimi-k3-fast' | 'morph/morph-v3-fast' | 'morph/morph-v3-large' | 'nvidia/nemotron-3-nano-30b-a3b' | 'nvidia/nemotron-3-super-120b-a12b' | 'nvidia/nemotron-3-ultra-550b-a55b' | 'nvidia/nemotron-3.5-lightning' | 'nvidia/nemotron-nano-12b-v2-vl' | 'nvidia/nemotron-nano-9b-v2' | 'openai/gpt-3.5-turbo' | 'openai/gpt-4-turbo' | 'openai/gpt-4.1' | 'openai/gpt-4.1-fast' | 'openai/gpt-4.1-mini' | 'openai/gpt-4.1-mini-fast' | 'openai/gpt-4.1-nano' | 'openai/gpt-4.1-nano-fast' | 'openai/gpt-4o' | 'openai/gpt-4o-fast' | 'openai/gpt-4o-mini' | 'openai/gpt-4o-mini-fast' | 'openai/gpt-5' | 'openai/gpt-5-codex' | 'openai/gpt-5-fast' | 'openai/gpt-5-mini' | 'openai/gpt-5-mini-fast' | 'openai/gpt-5-nano' | 'openai/gpt-5-pro' | 'openai/gpt-5.1-codex' | 'openai/gpt-5.1-codex-max' | 'openai/gpt-5.1-codex-mini' | 'openai/gpt-5.1-thinking' | 'openai/gpt-5.1-thinking-fast' | 'openai/gpt-5.2' | 'openai/gpt-5.2-codex' | 'openai/gpt-5.2-fast' | 'openai/gpt-5.2-pro' | 'openai/gpt-5.3-codex' | 'openai/gpt-5.3-codex-fast' | 'openai/gpt-5.4' | 'openai/gpt-5.4-fast' | 'openai/gpt-5.4-mini' | 'openai/gpt-5.4-mini-fast' | 'openai/gpt-5.4-nano' | 'openai/gpt-5.4-pro' | 'openai/gpt-5.5' | 'openai/gpt-5.5-fast' | 'openai/gpt-5.5-pro' | 'openai/gpt-5.6-luna' | 'openai/gpt-5.6-luna-fast' | 'openai/gpt-5.6-sol' | 'openai/gpt-5.6-sol-fast' | 'openai/gpt-5.6-terra' | 'openai/gpt-5.6-terra-fast' | 'openai/gpt-6-astra' | 'openai/gpt-6-astra-fast' | 'openai/gpt-6-luna' | 'openai/gpt-6-luna-fast' | 'openai/gpt-6-sol' | 'openai/gpt-6-sol-fast' | 'openai/gpt-oss-120b' | 'openai/gpt-oss-20b' | 'openai/gpt-oss-safeguard-120b' | 'openai/gpt-oss-safeguard-20b' | 'openai/o1' | 'openai/o3' | 'openai/o3-fast' | 'openai/o3-mini' | 'openai/o3-pro' | 'openai/o4-mini' | 'openai/o4-mini-fast' | 'perplexity/sonar' | 'perplexity/sonar-pro' | 'perplexity/sonar-reasoning-pro' | 'poolside/laguna-s-2.1' | 'poolside/laguna-s-2.1-free' | 'quiverai/arrow-2' | 'quiverai/arrow-2-telos' | 'sakana/fugu-max' | 'sakana/fugu-ultra' | 'sakana/fugu-ultra-v2' | 'sakana/namazu' | 'spacexai/grok-4.1-fast-non-reasoning' | 'spacexai/grok-4.1-fast-reasoning' | 'spacexai/grok-4.20-multi-agent' | 'spacexai/grok-4.20-multi-agent-beta' | 'spacexai/grok-4.20-non-reasoning' | 'spacexai/grok-4.20-non-reasoning-beta' | 'spacexai/grok-4.20-reasoning' | 'spacexai/grok-4.20-reasoning-beta' | 'spacexai/grok-4.3' | 'spacexai/grok-4.5' | 'spacexai/grok-4.6' | 'spacexai/grok-4.7' | 'spacexai/grok-build-0.1' | 'stepfun/step-3.5-flash' | 'stepfun/step-3.7-flash' | 'stepfun/step-5-preview' | 'tencent/hy-mt2-lite' | 'tencent/hy-mt2-plus' | 'tencent/hy-mt2-pro' | 'tencent/hy3' | 'tencent/hy4-preview' | 'thinkingmachines/inkling' | 'thinkingmachines/inkling-small' | 'xiaomi/mimo-v2.5' | 'xiaomi/mimo-v2.5-pro' | 'xiaomi/mimo-v2.6-flash' | 'xiaomi/mimo-v2.6-pro' | 'xiaomi/mimo-v2.6-pro-ultraspeed' | 'zai/glm-4.5' | 'zai/glm-4.5-air' | 'zai/glm-4.5v' | 'zai/glm-4.6' | 'zai/glm-4.7' | 'zai/glm-4.7-flash' | 'zai/glm-4.7-flashx' | 'zai/glm-5' | 'zai/glm-5-turbo' | 'zai/glm-5.1' | 'zai/glm-5.2' | 'zai/glm-5.2-fast' | 'zai/glm-5.3' | 'zai/glm-5.3-fast' | 'zai/glm-5.3-flash' | 'zai/glm-5.3-flashx' | 'zai/glm-5v-turbo' | (string & {});
12
12
 
13
13
  /**
14
14
  * Shared WebSocket subprotocol contract for AI Gateway realtime and streaming
@@ -81,7 +81,7 @@ type GatewayRealtimeModelId = 'google/gemini-3.8-live' | 'google/gemini-3.8-live
81
81
 
82
82
  type GatewayRerankingModelId = 'cohere/rerank-v3.5' | 'cohere/rerank-v4-fast' | 'cohere/rerank-v4-pro' | 'voyage/rerank-2.5' | 'voyage/rerank-2.5-lite' | (string & {});
83
83
 
84
- type GatewaySpeechModelId = 'fish-audio/s1' | 'fish-audio/s2-pro' | 'fish-audio/s2.1-pro' | 'openai/tts-1' | 'openai/tts-1-hd' | 'spacexai/grok-tts' | (string & {});
84
+ type GatewaySpeechModelId = 'fish-audio/s1' | 'fish-audio/s2-pro' | 'fish-audio/s2.1-pro' | 'google/gemini-3.8-flash-lite-tts' | 'google/gemini-3.8-flash-tts' | 'openai/tts-1' | 'openai/tts-1-hd' | 'spacexai/grok-tts' | (string & {});
85
85
 
86
86
  type GatewayTranscriptionModelId = 'fish-audio/transcribe-1' | 'google/gemini-3.5-transcribe' | 'google/gemini-3.5-transcribe-live' | 'openai/gpt-4o-mini-transcribe' | 'openai/gpt-4o-transcribe' | 'openai/gpt-realtime-whisper' | 'openai/whisper-1' | 'spacexai/grok-stt' | (string & {});
87
87
 
@@ -1120,11 +1120,17 @@ type GatewayProviderOptions = {
1120
1120
  /** Filter to providers that do not train on prompt data. */
1121
1121
  disallowPromptTraining?: boolean;
1122
1122
  /**
1123
- * Restrict routing to models that have all of the given capabilities.
1124
- * Currently supports `'implicit-caching'`, `'reasoning'`, `'tool-use'`, and
1125
- * `'vision'` (image input).
1126
- */
1127
- has?: Array<'implicit-caching' | 'reasoning' | 'tool-use' | 'vision'>;
1123
+ * Restrict routing to provider models that satisfy every given entry.
1124
+ *
1125
+ * Entries are capability tags (`'implicit-caching'`, `'reasoning'`,
1126
+ * `'tool-use'`, `'vision'`) or weight-format conditions: `'quantization:fp8'`
1127
+ * requires the serving provider to report that weight format,
1128
+ * `'!quantization:fp8'` excludes it (providers with no recorded format still
1129
+ * pass an exclusion). Format values are an open space but must match
1130
+ * `[a-zA-Z0-9._-]{1,32}` and compare case-insensitively; unknown capability
1131
+ * names are rejected by the Gateway with a 400.
1132
+ */
1133
+ has?: Array<'implicit-caching' | 'reasoning' | 'tool-use' | 'vision' | `quantization:${string}` | `!quantization:${string}`>;
1128
1134
  /**
1129
1135
  * Idempotency key for `experimental_startBatch`: retries with the same
1130
1136
  * key replay the original batch instead of creating a duplicate.
package/dist/index.js CHANGED
@@ -3662,7 +3662,7 @@ async function getVercelRequestId() {
3662
3662
  }
3663
3663
 
3664
3664
  // src/version.ts
3665
- var VERSION = true ? "4.0.90" : "0.0.0-test";
3665
+ var VERSION = true ? "4.0.92" : "0.0.0-test";
3666
3666
 
3667
3667
  // src/gateway-provider.ts
3668
3668
  var AI_GATEWAY_PROTOCOL_VERSION = "0.0.1";
package/docs/00-ai-gateway.mdx CHANGED
@@ -29,7 +29,7 @@ For most use cases, you can use the AI Gateway directly with a model string:
29
29
  import { generateText } from 'ai';
30
30
 
31
31
  const { text } = await generateText({
32
- model: 'openai/gpt-5.4',
32
+ model: 'openai/gpt-6-astra',
33
33
  prompt: 'Hello world',
34
34
  });
35
35
  ```
@@ -39,7 +39,7 @@ const { text } = await generateText({
39
39
  import { generateText, gateway } from 'ai';
40
40
 
41
41
  const { text } = await generateText({
42
- model: gateway('openai/gpt-5.4'),
42
+ model: gateway('openai/gpt-6-astra'),
43
43
  prompt: 'Hello world',
44
44
  });
45
45
  ```
@@ -200,7 +200,7 @@ You can create language models using a provider instance. The first argument is
200
200
  import { generateText } from 'ai';
201
201
 
202
202
  const { text } = await generateText({
203
- model: 'openai/gpt-5.4',
203
+ model: 'openai/gpt-6-astra',
204
204
  prompt: 'Explain quantum computing in simple terms',
205
205
  });
206
206
  ```
@@ -586,7 +586,7 @@ availableModels.models.forEach(model => {
586
586
 
587
587
  // Use any discovered model with plain string
588
588
  const { text } = await generateText({
589
- model: availableModels.models[0].id, // e.g., 'openai/gpt-5.4'
589
+ model: availableModels.models[0].id, // e.g., 'openai/gpt-6-astra'
590
590
  prompt: 'Hello world',
591
591
  });
592
592
  ```
@@ -620,7 +620,7 @@ import { gateway, generateText } from 'ai';
620
620
 
621
621
  // Make a request
622
622
  const result = await generateText({
623
- model: gateway('anthropic/claude-sonnet-4'),
623
+ model: gateway('anthropic/claude-sonnet-5'),
624
624
  prompt: 'Explain quantum entanglement briefly',
625
625
  });
626
626
 
@@ -643,7 +643,7 @@ With `streamText`, you can capture the generation ID from the first chunk via `s
643
643
  import { gateway, streamText } from 'ai';
644
644
 
645
645
  const result = streamText({
646
- model: gateway('anthropic/claude-sonnet-4'),
646
+ model: gateway('anthropic/claude-sonnet-5'),
647
647
  prompt: 'Explain quantum entanglement briefly',
648
648
  });
649
649
 
@@ -673,7 +673,7 @@ a tool is running:
673
673
  import { gateway, generateText } from 'ai';
674
674
 
675
675
  await generateText({
676
- model: gateway('anthropic/claude-sonnet-4'),
676
+ model: gateway('anthropic/claude-sonnet-5'),
677
677
  prompt: 'Explain quantum entanglement briefly',
678
678
  onLanguageModelCallEnd({ providerMetadata }) {
679
679
  const generationId = providerMetadata?.gateway?.generationId as
@@ -720,7 +720,7 @@ It returns a `GatewayGenerationInfo` object with the following fields:
720
720
  import { generateText } from 'ai';
721
721
 
722
722
  const { text } = await generateText({
723
- model: 'anthropic/claude-sonnet-4.6',
723
+ model: 'anthropic/claude-sonnet-5',
724
724
  prompt: 'Write a haiku about programming',
725
725
  });
726
726
 
@@ -733,7 +733,7 @@ console.log(text);
733
733
  import { streamText } from 'ai';
734
734
 
735
735
  const { textStream } = await streamText({
736
- model: 'openai/gpt-5.4',
736
+ model: 'openai/gpt-6-astra',
737
737
  prompt: 'Explain the benefits of serverless architecture',
738
738
  });
739
739
 
@@ -749,7 +749,7 @@ import { generateText, tool } from 'ai';
749
749
  import { z } from 'zod';
750
750
 
751
751
  const { text } = await generateText({
752
- model: 'xai/grok-4.7',
752
+ model: 'spacexai/grok-4.7',
753
753
  prompt: 'What is the weather like in San Francisco?',
754
754
  tools: {
755
755
  getWeather: tool({
@@ -775,7 +775,7 @@ import { generateText, isStepCount } from 'ai';
775
775
  import { openai } from '@ai-sdk/openai';
776
776
 
777
777
  const result = await generateText({
778
- model: 'openai/gpt-5.4-mini',
778
+ model: 'openai/gpt-6-luna',
779
779
  prompt: 'What is the Vercel AI Gateway?',
780
780
  stopWhen: isStepCount(10),
781
781
  tools: {
@@ -1207,7 +1207,7 @@ import type { GatewayProviderOptions } from '@ai-sdk/gateway';
1207
1207
  import { generateText } from 'ai';
1208
1208
 
1209
1209
  const { text } = await generateText({
1210
- model: 'openai/gpt-5.4',
1210
+ model: 'openai/gpt-6-astra',
1211
1211
  prompt: 'Summarize this document...',
1212
1212
  providerOptions: {
1213
1213
  gateway: {
@@ -1249,7 +1249,7 @@ The `getSpendReport()` method accepts the following parameters:
1249
1249
  - **groupBy** _string_ - Aggregation dimension: `'day'` (default), `'user'`, `'model'`, `'tag'`, `'provider'`, or `'credential_type'`
1250
1250
  - **datePart** _string_ - Time granularity when `groupBy` is `'day'`: `'day'` or `'hour'`
1251
1251
  - **userId** _string_ - Filter to a specific user
1252
- - **model** _string_ - Filter to a specific model (e.g. `'anthropic/claude-sonnet-4.5'`)
1252
+ - **model** _string_ - Filter to a specific model (e.g. `'anthropic/claude-sonnet-5'`)
1253
1253
  - **provider** _string_ - Filter to a specific provider (e.g. `'anthropic'`)
1254
1254
  - **credentialType** _string_ - Filter by `'byok'` or `'system'` credentials
1255
1255
  - **tags** _string[]_ - Filter to requests matching these tags
@@ -1308,7 +1308,7 @@ import type { GatewayProviderOptions } from '@ai-sdk/gateway';
1308
1308
  import { generateText } from 'ai';
1309
1309
 
1310
1310
  const { text } = await generateText({
1311
- model: 'anthropic/claude-sonnet-4.6',
1311
+ model: 'anthropic/claude-sonnet-5',
1312
1312
  prompt: 'Explain quantum computing',
1313
1313
  providerOptions: {
1314
1314
  gateway: {
@@ -1350,7 +1350,7 @@ The following gateway provider options are available:
1350
1350
 
1351
1351
  Specifies fallback models to use when the primary model fails or is unavailable. The gateway will try the primary model first (specified in the `model` parameter), then try each model in this array in order until one succeeds.
1352
1352
 
1353
- Example: `models: ['openai/gpt-5.4-nano', 'gemini-3-flash-preview']` will try the fallback models in order if the primary model fails.
1353
+ Example: `models: ['openai/gpt-5.4-nano', 'google/gemini-3.8-flash']` will try the fallback models in order if the primary model fails.
1354
1354
 
1355
1355
  - **user** _string_
1356
1356
 
@@ -1390,7 +1390,7 @@ The following gateway provider options are available:
1390
1390
 
1391
1391
  The unique identifier for the entity against which quota is tracked. Used for quota management and enforcement purposes.
1392
1392
 
1393
- - **has** _Array&lt;'implicit-caching' | 'reasoning' | 'tool-use' | 'vision'&gt;_
1393
+ - **has** _Array&lt;'implicit-caching' | 'reasoning' | 'tool-use' | 'vision' | `quantization:${string}` | `!quantization:${string}`&gt;_
1394
1394
 
1395
1395
  Restricts routing to provider models that have all of the specified capabilities. Applies to both BYOK and system credentials, since the capability is a property of the model rather than the credential. If no provider model for the requested model satisfies the capabilities, the request fails. Unsupported values are rejected.
1396
1396
 
@@ -1400,7 +1400,9 @@ The following gateway provider options are available:
1400
1400
  - `'tool-use'` — models that support tool calling.
1401
1401
  - `'vision'` — models that accept image input.
1402
1402
 
1403
- Example: `has: ['implicit-caching']` will only route to models that support implicit caching. Example: `has: ['vision']` will only route to models that accept image input. Example: `has: ['tool-use']` will only route to models that support tool calling.
1403
+ Weight-format conditions route on the serving provider's recorded weight format: `'quantization:fp8'` requires a provider serving fp8 weights, while `'!quantization:fp8'` excludes fp8 providers. A negated condition also matches providers whose weight format is not recorded, so exclusions remain usable while the catalog is populated.
1404
+
1405
+ Example: `has: ['implicit-caching']` will only route to models that support implicit caching. Example: `has: ['vision']` will only route to models that accept image input. Example: `has: ['tool-use']` will only route to models that support tool calling. Example: `has: ['!quantization:fp8']` will only route to providers that do not serve fp8 weights.
1404
1406
 
1405
1407
  - **providerTimeouts** _object_
1406
1408
 
@@ -1426,7 +1428,7 @@ The following gateway provider options are available:
1426
1428
  import { generateText } from 'ai';
1427
1429
 
1428
1430
  const { text } = await generateText({
1429
- model: 'openai/gpt-5-mini',
1431
+ model: 'openai/gpt-6-luna',
1430
1432
  prompt: 'Hello',
1431
1433
  providerOptions: {
1432
1434
  gateway: {
@@ -1443,7 +1445,7 @@ import type { GatewayProviderOptions } from '@ai-sdk/gateway';
1443
1445
  import { generateText } from 'ai';
1444
1446
 
1445
1447
  const { text } = await generateText({
1446
- model: 'anthropic/claude-sonnet-4.6',
1448
+ model: 'anthropic/claude-sonnet-5',
1447
1449
  prompt: 'Write a haiku about programming',
1448
1450
  providerOptions: {
1449
1451
  gateway: {
@@ -1463,11 +1465,11 @@ import type { GatewayProviderOptions } from '@ai-sdk/gateway';
1463
1465
  import { generateText } from 'ai';
1464
1466
 
1465
1467
  const { text } = await generateText({
1466
- model: 'openai/gpt-5.4', // Primary model
1468
+ model: 'openai/gpt-6-astra', // Primary model
1467
1469
  prompt: 'Write a TypeScript haiku',
1468
1470
  providerOptions: {
1469
1471
  gateway: {
1470
- models: ['openai/gpt-5.4-nano', 'gemini-3-flash-preview'], // Fallback models
1472
+ models: ['openai/gpt-5.4-nano', 'gemini-3.8-flash'], // Fallback models
1471
1473
  } satisfies GatewayProviderOptions,
1472
1474
  },
1473
1475
  });
@@ -1488,7 +1490,7 @@ import type { GatewayProviderOptions } from '@ai-sdk/gateway';
1488
1490
  import { generateText } from 'ai';
1489
1491
 
1490
1492
  const { text } = await generateText({
1491
- model: 'anthropic/claude-sonnet-4.6',
1493
+ model: 'anthropic/claude-sonnet-5',
1492
1494
  prompt: 'Analyze this sensitive document...',
1493
1495
  providerOptions: {
1494
1496
  gateway: {
@@ -1507,7 +1509,7 @@ import type { GatewayProviderOptions } from '@ai-sdk/gateway';
1507
1509
  import { generateText } from 'ai';
1508
1510
 
1509
1511
  const { text } = await generateText({
1510
- model: 'anthropic/claude-sonnet-4.6',
1512
+ model: 'anthropic/claude-sonnet-5',
1511
1513
  prompt: 'Analyze this proprietary business data...',
1512
1514
  providerOptions: {
1513
1515
  gateway: {
@@ -1526,7 +1528,7 @@ import type { GatewayProviderOptions } from '@ai-sdk/gateway';
1526
1528
  import { generateText } from 'ai';
1527
1529
 
1528
1530
  const { text } = await generateText({
1529
- model: 'anthropic/claude-sonnet-4.6',
1531
+ model: 'anthropic/claude-sonnet-5',
1530
1532
  prompt: 'Summarize this report...',
1531
1533
  providerOptions: {
1532
1534
  gateway: {
@@ -1538,14 +1540,14 @@ const { text } = await generateText({
1538
1540
 
1539
1541
  #### Filtering by Model Capability
1540
1542
 
1541
- Set `has` to restrict routing to provider models that have the specified capabilities. This applies to both BYOK and system credentials, since the capability is a property of the model rather than the credential. `'implicit-caching'` limits routing to models that perform automatic prompt caching, `'vision'` limits routing to models that accept image input, `'reasoning'` limits routing to models that support reasoning, and `'tool-use'` limits routing to models that support tool calling. If no provider model for the requested model satisfies the capabilities, the request fails.
1543
+ Set `has` to restrict routing to provider models that have the specified capabilities. This applies to both BYOK and system credentials, since the capability is a property of the model rather than the credential. `'implicit-caching'` limits routing to models that perform automatic prompt caching, `'vision'` limits routing to models that accept image input, `'reasoning'` limits routing to models that support reasoning, and `'tool-use'` limits routing to models that support tool calling. Weight-format conditions (`'quantization:fp8'` to require, `'!quantization:fp8'` to exclude) limit routing by the serving provider's recorded weight format. If no provider model for the requested model satisfies the capabilities, the request fails.
1542
1544
 
1543
1545
  ```ts
1544
1546
  import type { GatewayProviderOptions } from '@ai-sdk/gateway';
1545
1547
  import { generateText } from 'ai';
1546
1548
 
1547
1549
  const { text } = await generateText({
1548
- model: 'openai/gpt-5.5',
1550
+ model: 'openai/gpt-6-astra',
1549
1551
  prompt: 'Summarize this report...',
1550
1552
  providerOptions: {
1551
1553
  gateway: {
@@ -1564,7 +1566,7 @@ import { generateText } from 'ai';
1564
1566
  import fs from 'node:fs';
1565
1567
 
1566
1568
  const { text } = await generateText({
1567
- model: 'anthropic/claude-sonnet-4.6',
1569
+ model: 'anthropic/claude-sonnet-5',
1568
1570
  messages: [
1569
1571
  {
1570
1572
  role: 'user',
package/package.json CHANGED
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "name": "@ai-sdk/gateway",
3
3
  "private": false,
4
- "version": "4.0.90",
4
+ "version": "4.0.92",
5
5
  "type": "module",
6
6
  "license": "Apache-2.0",
7
7
  "sideEffects": false,
@@ -31,7 +31,7 @@
31
31
  },
32
32
  "dependencies": {
33
33
  "@ai-sdk/provider": "4.0.18",
34
- "@ai-sdk/provider-utils": "5.0.46",
34
+ "@ai-sdk/provider-utils": "5.0.47",
35
35
  "@vercel/oidc": "3.2.0"
36
36
  },
37
37
  "devDependencies": {
package/src/gateway-image-model-settings.ts CHANGED
@@ -27,6 +27,7 @@ export type GatewayImageModelId =
27
27
  | 'recraft/recraft-v4'
28
28
  | 'recraft/recraft-v4-pro'
29
29
  | 'recraft/recraft-v4.1'
30
+ | 'recraft/recraft-v4.1-flash'
30
31
  | 'recraft/recraft-v4.1-pro'
31
32
  | 'recraft/recraft-v4.1-utility'
32
33
  | 'recraft/recraft-v4.1-utility-pro'
package/src/gateway-language-model-settings.ts CHANGED
@@ -29,6 +29,7 @@ export type GatewayModelId =
29
29
  | 'alibaba/qwen3.8-flash'
30
30
  | 'alibaba/qwen3.8-max'
31
31
  | 'alibaba/qwen3.8-max-0902'
32
+ | 'alibaba/qwen3.8-max-prime'
32
33
  | 'alibaba/qwen3.8-omni-flash'
33
34
  | 'amazon/nova-2-lite'
34
35
  | 'amazon/nova-lite'
@@ -47,6 +48,7 @@ export type GatewayModelId =
47
48
  | 'anthropic/claude-opus-5'
48
49
  | 'anthropic/claude-opus-5-fast'
49
50
  | 'anthropic/claude-opus-5.5'
51
+ | 'anthropic/claude-opus-5.5-fast'
50
52
  | 'anthropic/claude-sonnet-4'
51
53
  | 'anthropic/claude-sonnet-4.5'
52
54
  | 'anthropic/claude-sonnet-4.6'
@@ -67,6 +69,7 @@ export type GatewayModelId =
67
69
  | 'deepseek/deepseek-v4-pro'
68
70
  | 'deepseek/deepseek-v4-pro-0813'
69
71
  | 'deepseek/deepseek-v4.1-flash'
72
+ | 'fireworks/ember-1'
70
73
  | 'google/gemini-2.5-flash'
71
74
  | 'google/gemini-2.5-flash-image'
72
75
  | 'google/gemini-2.5-flash-lite'
@@ -95,7 +98,6 @@ export type GatewayModelId =
95
98
  | 'inclusionai/ling-3.0-flash-sante'
96
99
  | 'inclusionai/ling-3.0-flash-sante-free'
97
100
  | 'inclusionai/ling-3.0-flash-vl'
98
- | 'inclusionai/ling-3.0-flash-vl-free'
99
101
  | 'inference-net/schematron-v2-small'
100
102
  | 'inference-net/schematron-v2-turbo'
101
103
  | 'interfaze/interfaze-beta'
package/src/gateway-provider-options.ts CHANGED
@@ -16,11 +16,24 @@ export type GatewayProviderOptions = {
16
16
  disallowPromptTraining?: boolean;
17
17
 
18
18
  /**
19
- * Restrict routing to models that have all of the given capabilities.
20
- * Currently supports `'implicit-caching'`, `'reasoning'`, `'tool-use'`, and
21
- * `'vision'` (image input).
19
+ * Restrict routing to provider models that satisfy every given entry.
20
+ *
21
+ * Entries are capability tags (`'implicit-caching'`, `'reasoning'`,
22
+ * `'tool-use'`, `'vision'`) or weight-format conditions: `'quantization:fp8'`
23
+ * requires the serving provider to report that weight format,
24
+ * `'!quantization:fp8'` excludes it (providers with no recorded format still
25
+ * pass an exclusion). Format values are an open space but must match
26
+ * `[a-zA-Z0-9._-]{1,32}` and compare case-insensitively; unknown capability
27
+ * names are rejected by the Gateway with a 400.
22
28
  */
23
- has?: Array<'implicit-caching' | 'reasoning' | 'tool-use' | 'vision'>;
29
+ has?: Array<
30
+ | 'implicit-caching'
31
+ | 'reasoning'
32
+ | 'tool-use'
33
+ | 'vision'
34
+ | `quantization:${string}`
35
+ | `!quantization:${string}`
36
+ >;
24
37
 
25
38
  /**
26
39
  * Idempotency key for `experimental_startBatch`: retries with the same
package/src/gateway-speech-model-settings.ts CHANGED
@@ -2,6 +2,8 @@ export type GatewaySpeechModelId =
2
2
  | 'fish-audio/s1'
3
3
  | 'fish-audio/s2-pro'
4
4
  | 'fish-audio/s2.1-pro'
5
+ | 'google/gemini-3.8-flash-lite-tts'
6
+ | 'google/gemini-3.8-flash-tts'
5
7
  | 'openai/tts-1'
6
8
  | 'openai/tts-1-hd'
7
9
  | 'spacexai/grok-tts'