CoreWeave

Data handling can vary by model and deployment. The privacy badge above shows the more conservative of the provider-wide fallback and the policies for routes on this page; each model page shows the policy for that specific route.

Models served by CoreWeave

Model name

30 models

Model	Context	Input /M	Output /M	Cache read /M	Cache hit	TTFT	Throughput
DeepSeek V3.1 FP8 deepseek-ai/DeepSeek-V3.1:coreweave	161k	$0.578/M	$1.73/M	$0.578/M	N/A	0.2s	78 tps
DeepSeek V4 Flash 0731 FP8 deepseek/deepseek-v4-flash-0731:coreweave	262k	$0.136/M	$0.294/M	$0.0735/M	N/A	0.8s	51 tps
DeepSeek V4 Flash 0731 (Thinking) FP8 deepseek/deepseek-v4-flash-0731:thinking:coreweave	262k	$0.136/M	$0.294/M	$0.0735/M	N/A	0.8s	51 tps
DeepSeek V4 Pro 0813 FP8 deepseek/deepseek-v4-pro-0813:coreweave	1.0M	$1.38/M	$4.16/M	$0.0462/M	N/A	0.8s	100 tps
DeepSeek V4 Pro 0813 (Thinking) FP8 deepseek/deepseek-v4-pro-0813:thinking:coreweave	1.0M	$1.38/M	$4.16/M	$0.0462/M	30.6%	0.8s	100 tps
DeepSeek V4.1 Flash FP8 deepseek/deepseek-v4.1-flash:coreweave	1.0M	$0.136/M	$0.682/M	$0.0021/M	N/A	0.5s	206 tps
DeepSeek V4.1 Flash (Thinking) FP8 deepseek/deepseek-v4.1-flash:thinking:coreweave	1.0M	$0.136/M	$0.682/M	$0.0021/M	1.1%	0.5s	206 tps
Gemma 4 26B A4B BF16 google/gemma-4-26b-a4b-it:coreweave	262k	$0.105/M	$0.315/M	$0.0525/M	N/A	0.3s	38 tps
Gemma 4 26B A4B (Thinking) BF16 google/gemma-4-26b-a4b-it:thinking:coreweave	262k	$0.105/M	$0.315/M	$0.0525/M	N/A	0.3s	38 tps
Gemma 4 31B FP4 google/gemma-4-31b-it:coreweave	262k	$0.105/M	$0.357/M	$0.105/M	10.5%	0.9s	35 tps
Gemma 4 31B (Thinking) FP4 google/gemma-4-31b-it:thinking:coreweave	262k	$0.105/M	$0.357/M	$0.105/M	9.2%	0.9s	35 tps
GLM 5.2 FP4 z-ai/glm-5.2:coreweave	1.0M	$0.651/M	$2.54/M	$0.147/M	0%	0.5s	112 tps
GLM 5.2 (Thinking) FP4 z-ai/glm-5.2:thinking:coreweave	1.0M	$0.651/M	$2.54/M	$0.147/M	20.2%	0.5s	112 tps
GLM 5.3 FP8 z-ai/glm-5.3:coreweave	1.0M	$1.26/M	$4.20/M	$0.20/M	39.6%	0.6s	91 tps
GLM 5.3 (Thinking) FP8 z-ai/glm-5.3:thinking:coreweave	1.0M	$1.26/M	$4.20/M	$0.20/M	39.1%	0.6s	91 tps
GLM 5.3 Flash NVFP4 z-ai/glm-5.3-flash:coreweave	1.0M	$0.126/M	$0.399/M	$0.021/M	21.7%	0.6s	138 tps
GPT OSS 120B FP4 openai/gpt-oss-120b:coreweave	131k	$0.0315/M	$0.178/M	$0.0315/M	1.3%	0.3s	48 tps
GPT OSS 20B FP4 openai/gpt-oss-20b:coreweave	131k	$0.0315/M	$0.136/M	$0.0315/M	N/A	0.2s	91 tps
Granite 4.2 8B BF16 ibm-granite/granite-4.2-8b:coreweave	131k	$0.063/M	$0.157/M	$0.0158/M	N/A	0.2s	108 tps
Kimi K2.6 FP4 moonshotai/kimi-k2.6:coreweave	262k	$0.682/M	$3.58/M	$0.157/M	0%	0.4s	106 tps
Kimi K2.6 Thinking FP4 moonshotai/kimi-k2.6:thinking:coreweave	262k	$0.682/M	$3.58/M	$0.157/M	41.9%	0.4s	106 tps
Kimi K2.7 Code INT4 moonshotai/kimi-k2.7-code:coreweave	262k	$0.63/M	$3.15/M	$0.126/M	12.3%	1.1s	68.5 tps
Llama 3.1 8B BF16 meta-llama/llama-3.1-8b-instruct:coreweave	131k	$0.231/M	$0.231/M	$0.231/M	N/A	0.3s	135 tps
Llama 3.3 70B Instruct FP16 meta-llama/llama-3.3-70b-instruct:coreweave	128k	$0.746/M	$0.746/M	$0.746/M	N/A	0.1s	87 tps
MiniMax M3 FP4 minimax/minimax-m3:coreweave	262k	$0.242/M	$1.01/M	$0.0525/M	57.3%	0.5s	120 tps
MiniMax M3 Thinking FP4 minimax/minimax-m3:thinking:coreweave	262k	$0.242/M	$1.01/M	$0.0525/M	0%	0.5s	120 tps
Qwen 3.6 35B A3B FP8 Qwen/Qwen3.6-35B-A3B:coreweave	262k	$0.105/M	$0.998/M	$0.0525/M	N/A	0.5s	111 tps
Qwen 3.6 35B A3B Thinking FP8 Qwen/Qwen3.6-35B-A3B:thinking:coreweave	262k	$0.105/M	$0.998/M	$0.0525/M	N/A	0.5s	111 tps
Qwen3.8 27B FP8 qwen3.8-27b:coreweave	262k	$0.42/M	$3.15/M	$0.157/M	N/A	0.6s	41 tps
Qwen3.8 27B Thinking FP8 qwen3.8-27b:thinking:coreweave	262k	$0.42/M	$3.15/M	$0.157/M	N/A	0.6s	41 tps

DeepSeek V3.1

FP8

deepseek-ai/DeepSeek-V3.1:coreweave

Context

161k

Input /M

$0.578/M

Output /M

$1.73/M

Cache read /M

$0.578/M

Cache hit

N/A

TTFT

0.2s

Throughput

78 tps

DeepSeek V4 Flash 0731

FP8

deepseek/deepseek-v4-flash-0731:coreweave

Context

262k

Input /M

$0.136/M

Output /M

$0.294/M

Cache read /M

$0.0735/M

Cache hit

N/A

TTFT

0.8s

Throughput

51 tps

DeepSeek V4 Flash 0731 (Thinking)

FP8

deepseek/deepseek-v4-flash-0731:thinking:coreweave

Context

262k

Input /M

$0.136/M

Output /M

$0.294/M

Cache read /M

$0.0735/M

Cache hit

N/A

TTFT

0.8s

Throughput

51 tps

DeepSeek V4 Pro 0813

FP8

deepseek/deepseek-v4-pro-0813:coreweave

Context

1.0M

Input /M

$1.38/M

Output /M

$4.16/M

Cache read /M

$0.0462/M

Cache hit

N/A

TTFT

0.8s

Throughput

100 tps

DeepSeek V4 Pro 0813 (Thinking)

FP8

deepseek/deepseek-v4-pro-0813:thinking:coreweave

Context

1.0M

Input /M

$1.38/M

Output /M

$4.16/M

Cache read /M

$0.0462/M

Cache hit

30.6%

TTFT

0.8s

Throughput

100 tps

DeepSeek V4.1 Flash

FP8

deepseek/deepseek-v4.1-flash:coreweave

Context

1.0M

Input /M

$0.136/M

Output /M

$0.682/M

Cache read /M

$0.0021/M

Cache hit

N/A

TTFT

0.5s

Throughput

206 tps

DeepSeek V4.1 Flash (Thinking)

FP8

deepseek/deepseek-v4.1-flash:thinking:coreweave

Context

1.0M

Input /M

$0.136/M

Output /M

$0.682/M

Cache read /M

$0.0021/M

Cache hit

1.1%

TTFT

0.5s

Throughput

206 tps

Gemma 4 26B A4B

BF16

google/gemma-4-26b-a4b-it:coreweave

Context

262k

Input /M

$0.105/M

Output /M

$0.315/M

Cache read /M

$0.0525/M

Cache hit

N/A

TTFT

0.3s

Throughput

38 tps

Gemma 4 26B A4B (Thinking)

BF16

google/gemma-4-26b-a4b-it:thinking:coreweave

Context

262k

Input /M

$0.105/M

Output /M

$0.315/M

Cache read /M

$0.0525/M

Cache hit

N/A

TTFT

0.3s

Throughput

38 tps

Gemma 4 31B

FP4

google/gemma-4-31b-it:coreweave

Context

262k

Input /M

$0.105/M

Output /M

$0.357/M

Cache read /M

$0.105/M

Cache hit

10.5%

TTFT

0.9s

Throughput

35 tps

Gemma 4 31B (Thinking)

FP4

google/gemma-4-31b-it:thinking:coreweave

Context

262k

Input /M

$0.105/M

Output /M

$0.357/M

Cache read /M

$0.105/M

Cache hit

9.2%

TTFT

0.9s

Throughput

35 tps

GLM 5.2

FP4

z-ai/glm-5.2:coreweave

Context

1.0M

Input /M

$0.651/M

Output /M

$2.54/M

Cache read /M

$0.147/M

Cache hit

0%

TTFT

0.5s

Throughput

112 tps

GLM 5.2 (Thinking)

FP4

z-ai/glm-5.2:thinking:coreweave

Context

1.0M

Input /M

$0.651/M

Output /M

$2.54/M

Cache read /M

$0.147/M

Cache hit

20.2%

TTFT

0.5s

Throughput

112 tps

GLM 5.3

FP8

z-ai/glm-5.3:coreweave

Context

1.0M

Input /M

$1.26/M

Output /M

$4.20/M

Cache read /M

$0.20/M

Cache hit

39.6%

TTFT

0.6s

Throughput

91 tps

GLM 5.3 (Thinking)

FP8

z-ai/glm-5.3:thinking:coreweave

Context

1.0M

Input /M

$1.26/M

Output /M

$4.20/M

Cache read /M

$0.20/M

Cache hit

39.1%

TTFT

0.6s

Throughput

91 tps

GLM 5.3 Flash

NVFP4

z-ai/glm-5.3-flash:coreweave

Context

1.0M

Input /M

$0.126/M

Output /M

$0.399/M

Cache read /M

$0.021/M

Cache hit

21.7%

TTFT

0.6s

Throughput

138 tps

GPT OSS 120B

FP4

openai/gpt-oss-120b:coreweave

Context

131k

Input /M

$0.0315/M

Output /M

$0.178/M

Cache read /M

$0.0315/M

Cache hit

1.3%

TTFT

0.3s

Throughput

48 tps

GPT OSS 20B

FP4

openai/gpt-oss-20b:coreweave

Context

131k

Input /M

$0.0315/M

Output /M

$0.136/M

Cache read /M

$0.0315/M

Cache hit

N/A

TTFT

0.2s

Throughput

91 tps

Granite 4.2 8B

BF16

ibm-granite/granite-4.2-8b:coreweave

Context

131k

Input /M

$0.063/M

Output /M

$0.157/M

Cache read /M

$0.0158/M

Cache hit

N/A

TTFT

0.2s

Throughput

108 tps

Kimi K2.6

FP4

moonshotai/kimi-k2.6:coreweave

Context

262k

Input /M

$0.682/M

Output /M

$3.58/M

Cache read /M

$0.157/M

Cache hit

0%

TTFT

0.4s

Throughput

106 tps

Kimi K2.6 Thinking

FP4

moonshotai/kimi-k2.6:thinking:coreweave

Context

262k

Input /M

$0.682/M

Output /M

$3.58/M

Cache read /M

$0.157/M

Cache hit

41.9%

TTFT

0.4s

Throughput

106 tps

Kimi K2.7 Code

INT4

moonshotai/kimi-k2.7-code:coreweave

Context

262k

Input /M

$0.63/M

Output /M

$3.15/M

Cache read /M

$0.126/M

Cache hit

12.3%

TTFT

1.1s

Throughput

68.5 tps

Llama 3.1 8B

BF16

meta-llama/llama-3.1-8b-instruct:coreweave

Context

131k

Input /M

$0.231/M

Output /M

$0.231/M

Cache read /M

$0.231/M

Cache hit

N/A

TTFT

0.3s

Throughput

135 tps

Llama 3.3 70B Instruct

FP16

meta-llama/llama-3.3-70b-instruct:coreweave

Context

128k

Input /M

$0.746/M

Output /M

$0.746/M

Cache read /M

$0.746/M

Cache hit

N/A

TTFT

0.1s

Throughput

87 tps

MiniMax M3

FP4

minimax/minimax-m3:coreweave

Context

262k

Input /M

$0.242/M

Output /M

$1.01/M

Cache read /M

$0.0525/M

Cache hit

57.3%

TTFT

0.5s

Throughput

120 tps

MiniMax M3 Thinking

FP4

minimax/minimax-m3:thinking:coreweave

Context

262k

Input /M

$0.242/M

Output /M

$1.01/M

Cache read /M

$0.0525/M

Cache hit

0%

TTFT

0.5s

Throughput

120 tps

Qwen 3.6 35B A3B

FP8

Qwen/Qwen3.6-35B-A3B:coreweave

Context

262k

Input /M

$0.105/M

Output /M

$0.998/M

Cache read /M

$0.0525/M

Cache hit

N/A

TTFT

0.5s

Throughput

111 tps

Qwen 3.6 35B A3B Thinking

FP8

Qwen/Qwen3.6-35B-A3B:thinking:coreweave

Context

262k

Input /M

$0.105/M

Output /M

$0.998/M

Cache read /M

$0.0525/M

Cache hit

N/A

TTFT

0.5s

Throughput

111 tps

Qwen3.8 27B

FP8

qwen3.8-27b:coreweave

Context

262k

Input /M

$0.42/M

Output /M

$3.15/M

Cache read /M

$0.157/M

Cache hit

N/A

TTFT

0.6s

Throughput

41 tps

Qwen3.8 27B Thinking

FP8

qwen3.8-27b:thinking:coreweave

Context

262k

Input /M

$0.42/M

Output /M

$3.15/M

Cache read /M

$0.157/M

Cache hit

N/A

TTFT

0.6s

Throughput

41 tps

Prices are per million tokens. TTFT is the observed time to first token. Selectable routes include the applicable provider-selection markup; fixed routes show the standard model price. Availability and pricing refresh continuously; each model page shows the live provider comparison.

Using CoreWeave via the API

Call a model served by CoreWeave using its displayed model ID, including the provider suffix when selection is supported.

curl https://nano-gpt.com/api/v1/chat/completions \
  -H "Authorization: Bearer $NANOGPT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-ai/DeepSeek-V3.1:coreweave",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

For models with provider selection, append :coreweave to the model ID, set "provider": "coreweave" in the request body, or send an X-Provider: coreweave header. Explicit provider selection adds a route-specific markup over that provider's base price; the prices in the table above already include it. Models without provider selection use the displayed model ID without a provider suffix or selection surcharge.

See the API documentation for provider preferences, price-aware routing, and error behavior.