Fireworks: Models Intelligence, Performance & Price
This analysis is intended to support you in choosing the best model provided by Fireworks for your use-case.
Most Intelligent
Updated
GLM-5.3-Flash
Qwen3.8 2.4T A95B
DeepSeek V4 Pro 0813 (max)DeepSeek V4 Flash Vision (max)DeepSeek V4 Flash 0731 (max)Intelligence index
Total 13 models
Fastest
Nemotron 3.5 Lightning (BF16)DeepSeek V4 Flash 0731 (max)GLM-5.2 (max)GLM-5.3-FlashNemotron 3 UltraOutput speed
Total 13 models
Lowest Price
Nemotron 3.5 Lightning (BF16)GLM-5.3-FlashDeepSeek V4 Flash Vision (max)DeepSeek V4 Flash 0731 (max)
MiniMax-M3Blended price (per 1M tokens)
Total 13 models
Fireworks offers 13 models, each with different intelligence, performance, and pricing characteristics. Below is a comparison of the key metrics across models.
- For intelligence, the top models on Fireworks are GLM-5.3-Flash (42), Qwen3.8 2.4T A95B (40), and DeepSeek V4 Pro 0813 (max) (36).
- For output speed, the fastest models are Nemotron 3.5 Lightning (BF16) (492 t/s), DeepSeek V4 Flash 0731 (max) (256 t/s), and GLM-5.2 (max) (216 t/s). Speed varies significantly across models, with a 188% difference between the fastest and slowest.
- For latency, Kimi K2.6 (Non-reasoning) (0.85s), Nemotron 3.5 Lightning (BF16) (4.54s), and DeepSeek V4 Flash 0731 (max) (8.42s) offer the lowest time to first answer token.
- For pricing, Nemotron 3.5 Lightning (BF16) ($0.04), GLM-5.3-Flash ($0.10), and DeepSeek V4 Flash Vision (max) ($0.11) offer the lowest blended prices per 1M tokens. Prices vary up to 6.0x across models.
- For context window size, GLM-5.3-Flash (1M), DeepSeek V4 Flash 0731 (max) (1M), and DeepSeek V4 Pro 0813 (max) (1M) support the largest context windows on Fireworks.
- Nemotron 3.5 Lightning (BF16) offers both the fastest output and best pricing, making it attractive for throughput-sensitive and cost-conscious applications. GLM-5.3-Flash leads in intelligence for tasks that require the highest quality.
Highlights
Intelligence Evaluations
Artificial Analysis Intelligence Index
Intelligence Evaluations
Agentic knowledge work, (Elo-500)/2000
Agentic real-world work tasks, (Elo-500)/2000
Agentic SaaS workflows
Agentic coding & terminal use
Coding
Reasoning & knowledge
Professional document reasoning, All-pass
Physics reasoning
Knowledge
1 - hallucination rate
Long context reasoning
Legal agentic work, criterion pass rate
Agentic business operations
Quantitative analysis on spreadsheets & documents
Instruction following
Agentic tool use
Long-horizon agentic tasks
Kubernetes incident root-cause analysis
Visual reasoning
Intelligence Index vs. Price
Context Window
Context Window
Pricing
Intelligence Index vs. Price
Performance Summary
Output Speed vs. Price
Speed
Measured by Output Speed (tokens per second)
Output Speed
Latency
Measured by Time (seconds) to First Token
Latency: Time To First Answer Token
End-to-End Response Time
Seconds to output 500 tokens, calculated based on time to first token, 'thinking' time for reasoning models, and output speed
End-to-End Response Time vs. Price
Further Analysis | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
GLM-5.3 (max) | 1.05M | Open | 45 | $2.15 | 216 | 0.86 | 12.43 | 9.25 | |||
 Kimi K3 (max) | 1.05M | Open | 44 | $5.42 | 111 | 0.98 | 23.41 | 17.94 | |||
Kimi K3 (max) (FAST) | 1.05M | Open | 44 | $5.80 | 111 | 0.96 | 23.54 | 18.06 | |||
GLM-5.3-Flash | 1.05M | Open | 42 | $0.30 | 179 | 1.36 | 15.36 | 11.20 | |||
Qwen3.8 2.4T A95B | 262k | Open | 40 | $1.87 | 81 | 1.11 | 31.89 | 24.62 | |||
GLM-5.2 (max) | 1M | Open | 39* | -- | 98 | 1.84 | 27.47 | 20.50 | |||
DeepSeek V4 Pro 0813 (max) | 1M | Open | 36 | $0.60 | 89 | 1.54 | 29.49 | 22.36 | |||
DeepSeek V4 Flash Vision (max) | 1M | Proprietary | 35 | $0.12 | 143 | 0.80 | 18.31 | 14.01 | |||
DeepSeek V4 Flash 0731 (max) | 1.05M | Open | 35 | $0.09 | 254 | 0.60 | 10.44 | 7.87 | |||
Kimi K2.6 | 262k | Open | 31* | -- | 67 | 0.74 | 74.89 | 66.66 | |||
MiniMax-M3 | 524k | Open | 30 | $0.50 | 139 | 1.07 | 19.11 | 14.44 | |||
GLM-5 | 203k | Open | 28* | -- | -- | -- | -- | -- | |||
Kimi K2.7 Code | 262k | Open | 26 | $0.52 | 103 | 0.72 | 27.34 | 21.74 | |||
Kimi K2.6 (Non-reasoning) | 262k | Open | 24* | -- | 74 | 0.78 | 7.54 | -- | |||
Nemotron 3 Ultra | 262k | Open | 23 | $0.22 | 155 | 0.64 | 18.51 | 14.65 | |||
GLM-5 (Non-reasoning) | 203k | Open | 22* | -- | -- | -- | -- | -- | |||
DeepSeek V3.2 | 164k | Open | 21* | -- | -- | -- | -- | -- | |||
 Muse Glimmer (high) | 131k | Open | 18 | $0.06 | 60 | 1.12 | 42.61 | 33.19 | |||
DeepSeek V3.2 (Non-reasoning) | 164k | Open | 16* | -- | -- | -- | -- | -- | |||
DeepSeek V3.1 (Non-reasoning) | 164k | Open | 14* | -- | -- | -- | -- | -- | |||
Nemotron 3.5 Lightning (BF16) | 262k | Open | 14 | $0.02 | 477 | 0.47 | 5.72 | 4.19 | |||
Qwen3 VL 235B A22B | 262k | Open | 10* | -- | -- | -- | -- | -- | |||
Qwen3 VL 30B A3B (Reasoning) | 262k | Open | 9* | -- | -- | -- | -- | -- | |||
Qwen3 VL 30B A3B | 262k | Open | 8* | -- | -- | -- | -- | -- | |||
Llama 3.3 70B | 131k | Open | 8* | -- | -- | -- | -- | -- | |||
Qwen3 30B | 262k | Open | 8* | -- | -- | -- | -- | -- | |||
Qwen3 8B (Non-reasoning) | 41k | Open | 6* | -- | -- | -- | -- | -- | |||
Key definitions
Frequently Asked Questions
Common questions about Fireworks
Fireworks offers 13 models that we track: GLM-5.3-Flash, Qwen3.8 2.4T A95B, DeepSeek V4 Pro 0813 (max), DeepSeek V4 Flash Vision (max), DeepSeek V4 Flash 0731 (max), GLM-5.2 (max), Kimi K2.6, MiniMax-M3, Kimi K2.7 Code, Kimi K2.6 (Non-reasoning), Nemotron 3 Ultra, Muse Glimmer (high), and Nemotron 3.5 Lightning (BF16).
The most intelligent model available on Fireworks is GLM-5.3-Flash with an Intelligence Index score of 42.
The fastest model on Fireworks by output speed is Nemotron 3.5 Lightning (BF16) at 492.1 tokens per second.
The model with the lowest time to first answer token on Fireworks is Kimi K2.6 (Non-reasoning) at 0.85s. Lower latency means faster initial response time.
The most affordable model on Fireworks by blended price is Nemotron 3.5 Lightning (BF16) at $0.04 per 1M tokens (7:2:1 cache hit/input/output ratio).
Prices on Fireworks vary up to 32x across models, from $0.04 per 1M tokens for Nemotron 3.5 Lightning (BF16) to $1.18 per 1M tokens for Qwen3.8 2.4T A95B.
Yes, Fireworks offers an OpenAI-compatible API, making it easy to switch from OpenAI or use existing OpenAI SDK integrations.
Yes, all 13 models on Fireworks support JSON mode for structured output.
Yes, all 13 models on Fireworks support function calling (tool use).
Yes, Fireworks offers 12 reasoning models: GLM-5.3-Flash, Qwen3.8 2.4T A95B, DeepSeek V4 Pro 0813 (max), DeepSeek V4 Flash Vision (max), DeepSeek V4 Flash 0731 (max), GLM-5.2 (max), Kimi K2.6, MiniMax-M3, Kimi K2.7 Code, Nemotron 3 Ultra, Muse Glimmer (high), and Nemotron 3.5 Lightning (BF16). Reasoning models use extended thinking to work through complex problems before providing an answer.
Yes, 12 of 13 models on Fireworks are open weight models: GLM-5.3-Flash, Qwen3.8 2.4T A95B, DeepSeek V4 Pro 0813 (max), DeepSeek V4 Flash 0731 (max), GLM-5.2 (max), Kimi K2.6, MiniMax-M3, Kimi K2.7 Code, Kimi K2.6 (Non-reasoning), Nemotron 3 Ultra, Muse Glimmer (high), and Nemotron 3.5 Lightning (BF16).
Yes, provider performance can vary over time due to infrastructure changes, load balancing, and updates. We continuously benchmark all providers and display historical performance trends in the "Over Time" charts.
When choosing a model on Fireworks, consider: intelligence (for quality-sensitive tasks), output speed (for throughput-intensive tasks), latency (for interactive applications requiring quick first responses), pricing (for cost-sensitive workloads), and features like context window size, JSON mode, or function calling support.