Fireworks AI logo

Fireworks AI

LLM 5.6/10

Fireworks AI provides fast, production-grade inference on open models (Llama, DeepSeek, Qwen, gpt-oss) with $1 in free credits for new accounts.

Fireworks AI scores 5.6/10 for free value in LLM. The main constraint: $1 in free credits on signup for serverless text and vision models; pay-as-you-go thereafter. Finite, one-time.

Category LLM Kind Fast hosted open-model LLM API Caveats Only $1 of starter credit, so the free window is brief. Card/top-up required to continue. Open-weight model catalog. Card required Yes Eligibility anyone

Free value score

5.6 / 10
Show the math
Value 5/10
value = 1 $ one-time
weighted score = 0.7 * 5/10
Model Quality 9.2/10
value = 51 artificial_analysis_score
weighted score = 0.2 * 9.2/10
Volume 2.4/10
value = 519481 tok/mo
score = 2.4/10
Request Rate 3/10
value = 10 req/min
weighted score = 0.1 * 3/10
Access 7/10
value = 7/10
score = 7/10

Key metrics

GLM-5.2 51 Best model
10 R/min Requests/min
$1 Credit
Yes Commercial use

Models

0 free · 100+ total
Model Family AA Access
GLM-5.2 GLM 51.1 paid
MiniMax-M3 MiniMax 44.4 paid
DeepSeek V4 Pro DeepSeek 44.3 paid
Kimi K2.6 Kimi 42.8 paid
DeepSeek V4 Flash DeepSeek 40.3 paid
GLM 5.1 GLM 40.2 paid
Kimi K2.5 Kimi 38.1 paid
DeepSeek V3.2 DeepSeek 24.7 paid
OpenAI gpt-oss-120b GPT 23.8 paid
DeepSeek V3.1 DeepSeek 21.1 paid
DeepSeek R1 DeepSeek 20.1 paid
Kimi K2 Thinking Kimi 19.4 paid
Llama 4 Maverick Llama 14.3 paid
Llama 4 Scout Llama 10 paid
Devstral Small 2 24B Instruct 2512 Mistral — paid
FLUX.1 [dev] FLUX — paid
FLUX.1 Kontext FLUX — paid
Gemma 3 12B Instruct Gemma — paid
KAT Coder Other — paid
Llama 3.3 70B Instruct Llama — paid
MiniMax 2.5 (M2.5) MiniMax — paid
MiniMax 2.7 (M2.7) MiniMax — paid
Mixtral 8x22B Instruct Mistral — paid
Mixtral 8x7B Instruct Mistral — paid
Molmo2 Other — paid
Nomic / GTE / Qwen3 embeddings Other — paid
OpenAI gpt-oss-20b GPT — paid
Qwen 3.6 Plus Qwen — paid
Qwen2.5-VL 32B Instruct Qwen — paid
Qwen3 235B A22B Qwen — paid
Qwen3 30B A3B Qwen — paid
Qwen3 Coder 480B A35B Qwen — paid
Qwen3 Omni Qwen — paid
Qwen3 VL 235B Qwen — paid

Showing first 0 of a larger catalog.

What you get

$1 free credit for new accounts

Optimized fast inference for agentic + structured output

Llama, DeepSeek, Qwen, gpt-oss and vision models

OpenAI-compatible API

Limitations

  • $1 one-time starter credit

    New accounts receive exactly $1 in free credits for serverless text and vision models; once exhausted there is no renewal or top-up from Fireworks.

  • Card required to continue

    After the $1 credit is consumed, a payment method must be added to continue using the API — the free credit alone is not sufficient for sustained use.

  • Evaluation window only

    At typical serverless token pricing, $1 covers only tens of thousands of tokens — enough for brief evaluation but not meaningful development or production use.

Sources

fireworks fast-inference open-weight finite-credit structured-output
Open service Source ← All LLM