[adarshsingh705]

Description

Adds a new provider file pricing/doubleword.json for Doubleword AI with the 18 currently listed models. Models scheduled for shutdown on 30 Oct 2026 are not included.

• Realtime rates → pay_as_you_go
• Batch (24h) rates → batch_config
• cache_write_input_token = 1.25× standard input. This is the explicit 5-minute cache write rate from the prompt-caching guide. Implicit caching never bills writes.
• lightonai/LightOnOCR-2-1B-bbox-soup has no Realtime price on the official page, so it only has batch_config.
• The Async tier is not included, to keep this PR small. I can add it under custom_pricing.execution_modes if you prefer.
• The default entry has zero prices and "currency": "USD", with the same structure as the default in anthropic.json.
• Model keys are the exact API model IDs from each model's docs page.

Note: the issue mentions a 1-hour SLA, but the pricing page currently lists only Realtime, Async and Batch (24h). The gateway has no doubleword provider yet, so this data is ready for when one is added.

• New model pricing
• Update existing pricing
• New model configuration
• Bug fix

Source Verification

Source Link: https://docs.doubleword.ai/batches/model-pricing
Per-model pages (model IDs): https://docs.doubleword.ai/inference-api/models
Cache write pricing: https://docs.doubleword.ai/inference-api/prompt-caching

Conversion (cents per token = $/1M ÷ 1,000,000 × 100)

Model	Tier	Input $/1M → ¢/token	Cache read $/1M → ¢/token	Output $/1M → ¢/token	Cache write (1.25× input) ¢/token
deepseek-ai/DeepSeek-V4.1-Flash-1017	Realtime	$0.15 → 0.000015	$0.01 → 0.000001	$0.6 → 0.00006	0.00001875
deepseek-ai/DeepSeek-V4.1-Flash-1017	Batch (24h)	$0.08 → 0.000008	$0.01 → 0.000001	$0.3 → 0.00003	—
meta/muse-spark-open	Realtime	$1.25 → 0.000125	$0.15 → 0.000015	$4.25 → 0.000425	0.00015625
meta/muse-spark-open	Batch (24h)	$1.25 → 0.000125	$0.15 → 0.000015	$4.25 → 0.000425	—
deepseek-ai/DeepSeek-V4.1-Flash	Realtime	$0.15 → 0.000015	$0.01 → 0.000001	$0.6 → 0.00006	0.00001875
deepseek-ai/DeepSeek-V4.1-Flash	Batch (24h)	$0.08 → 0.000008	$0.01 → 0.000001	$0.3 → 0.00003	—
zai-org/GLM-5.3-Flash	Realtime	$0.15 → 0.000015	$0.03 → 0.000003	$0.5 → 0.00005	0.00001875
zai-org/GLM-5.3-Flash	Batch (24h)	$0.08 → 0.000008	$0.02 → 0.000002	$0.25 → 0.000025	—
zai-org/GLM-5.3	Realtime	$1.4 → 0.00014	$0.28 → 0.000028	$4.4 → 0.00044	0.000175
zai-org/GLM-5.3	Batch (24h)	$0.7 → 0.00007	$0.14 → 0.000014	$2.2 → 0.00022	—
Qwen/Qwen3.8-27B-FP8	Realtime	$0.45 → 0.000045	$0.04 → 0.000004	$3 → 0.0003	0.00005625
Qwen/Qwen3.8-27B-FP8	Batch (24h)	$0.25 → 0.000025	$0.02 → 0.000002	$1.5 → 0.00015	—
deepseek-ai/DeepSeek-V4-Pro-0813	Realtime	$1.32 → 0.000132	$0.13 → 0.000013	$3.96 → 0.000396	0.000165
deepseek-ai/DeepSeek-V4-Pro-0813	Batch (24h)	$0.66 → 0.000066	$0.07 → 0.000007	$1.98 → 0.000198	—
meta-models/Muse-Glimmer-30B	Realtime	$0.11 → 0.000011	$0.01 → 0.000001	$0.32 → 0.000032	0.00001375
meta-models/Muse-Glimmer-30B	Batch (24h)	$0.05 → 0.000005	$0.01 → 0.000001	$0.15 → 0.000015	—
deepseek-ai/DeepSeek-V4-Flash-0731	Realtime	$0.09 → 0.000009	$0.02 → 0.000002	$0.18 → 0.000018	0.00001125
deepseek-ai/DeepSeek-V4-Flash-0731	Batch (24h)	$0.05 → 0.000005	$0.01 → 0.000001	$0.09 → 0.000009	—
moonshotai/kimi-k3	Realtime	$3 → 0.0003	$0.3 → 0.00003	$15 → 0.0015	0.000375
moonshotai/kimi-k3	Batch (24h)	$1.5 → 0.00015	$0.15 → 0.000015	$7.5 → 0.00075	—
zai-org/GLM-5.2-FP8	Realtime	$0.93 → 0.000093	$0.19 → 0.000019	$3 → 0.0003	0.00011625
zai-org/GLM-5.2-FP8	Batch (24h)	$0.47 → 0.000047	$0.09 → 0.000009	$1.5 → 0.00015	—
nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4	Realtime	$0.5 → 0.00005	$0.13 → 0.000013	$2.2 → 0.00022	0.0000625
nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4	Batch (24h)	$0.25 → 0.000025	$0.06 → 0.000006	$1.1 → 0.00011	—
Qwen/Qwen3.6-35B-A3B-FP8	Realtime	$0.14 → 0.000014	$0.14 → 0.000014	$1 → 0.0001	0.0000175
Qwen/Qwen3.6-35B-A3B-FP8	Batch (24h)	$0.07 → 0.000007	$0.07 → 0.000007	$0.5 → 0.00005	—
google/gemma-4-31B-it	Realtime	$0.12 → 0.000012	$0.12 → 0.000012	$0.35 → 0.000035	0.000015
google/gemma-4-31B-it	Batch (24h)	$0.06 → 0.000006	$0.06 → 0.000006	$0.18 → 0.000018	—
Qwen/Qwen3.5-35B-A3B-FP8-dottxt	Realtime	$0.35 → 0.000035	$0.2 → 0.00002	$1.2 → 0.00012	0.00004375
Qwen/Qwen3.5-35B-A3B-FP8-dottxt	Batch (24h)	$0.1 → 0.00001	$0.06 → 0.000006	$0.4 → 0.00004	—
lightonai/LightOnOCR-2-1B-bbox-soup	Realtime	not offered	—	—	—
lightonai/LightOnOCR-2-1B-bbox-soup	Batch (24h)	$0.05 → 0.000005	$0.05 → 0.000005	$0.05 → 0.000005	—
Qwen/Qwen3-VL-235B-A22B-Instruct-FP8	Realtime	$0.21 → 0.000021	$0.12 → 0.000012	$1.9 → 0.00019	0.00002625
Qwen/Qwen3-VL-235B-A22B-Instruct-FP8	Batch (24h)	$0.11 → 0.000011	$0.06 → 0.000006	$0.95 → 0.000095	—
Qwen/Qwen3-Embedding-8B	Realtime	$0.04 → 0.000004	$0.04 → 0.000004	$0 → 0	0.000005
Qwen/Qwen3-Embedding-8B	Batch (24h)	$0.02 → 0.000002	$0.02 → 0.000002	$0 → 0	—

Checklist

• I have validated the JSON using jq or an online validator
• I have verified that prices are in cents per token (not dollars)
• I have included the source link above
• I have signed the CLA (if first-time contributor)

Related Issue

Fixes #30