impulse

Qwen3 235B A22B Thinking 2507

Open weights

Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks.

API id qwen3-235b-a22b-thinking-2507Released 25 Jul 2025 (Jul 2025)
Input / 1M
$0.22
at Fireworks AI
Output / 1M
$0.88
at Fireworks AI
Context window
262K
tokens
Max output
262K
tokens per reply
Hosts
4
4 listings
Released
Jul 2025
25 Jul 2025

Pricing

List prices at Fireworks AI, in US dollars per million tokens.

TierInput / 1MOutput / 1M
Standard$0.22$0.88

What it costs a month

At Fireworks AI list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out$132
  • RAG over documents2,000 requests a day, 8K in / 0.6K out$137
  • Coding agent1,000 requests a day, 40K in / 2K out$317
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$739
Use your own numbers

Where to run it

4 hosts serve Qwen3 235B A22B Thinking 2507. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
Fireworks AIInference
qwen3-235b-a22b-thinking-2507
$0.22 / $0.88
DeepInfraInference
Qwen3-235B-A22B-Thinking-2507
$0.30 / $2.90
+147% vs headline
Novita AIInference
qwen3-235b-a22b-thinking-2507
$0.30 / $3.00
+153% vs headline
GMI CloudInference
Qwen3-235B-A22B-Thinking-2507-FP8
$0.35 / $4.20
+241% vs headline

Capabilities

Core

  • Streaming
  • Function calling
  • JSON mode
  • Structured outputs

Multimodal

  • Vision
  • Image input
  • Audio
  • Image generation
  • Video

Advanced

  • Reasoning
  • Tool calling
  • MCP
  • Prompt caching
  • Batch API
  • Embeddings
Takes in
Text
Produces
Text
Open weights
License: Apache-2.0
Runs in about 564.2 GB of GPU memory
Runtimes: Ollama, SGLang, llama.cpp, vLLM

Cheaper alternatives

Current models that cost less and keep most of what this one can do: tools, vision, reasoning, structured output, caching and enough context.

Same model, cheaper host

No host sells it for less than Fireworks AI.

Smaller Alibaba Qwen models

No cheaper Alibaba Qwen model keeps its key capabilities.

Other frontier labs

  • Claude Haiku 5.5
    Anthropic · on Anthropic
    $0.10 / $0.50
    Has every key capability
    48% less
  • GPT-6 Luna
    OpenAI · on OpenAI
    $0.10 / $0.50
    Has every key capability
    48% less
  • GPT-4o Mini
    OpenAI · on OpenAI
    $0.15 / $0.60
    Has every key capability
    32% less

Open-weight picks

  • Qwen3.5 397B A17B
    Alibaba Qwen · on AIHubMix
    $0.164 / $0.986
    Has every key capability
    4% less
  • Muse Glimmer 30B
    Meta · on EmpirioLabs AI
    $0.20 / $0.80
    Has every key capability
    9% less
  • Mistral Small 2603
    Mistral AI · on Mistral AI
    $0.15 / $0.60
    Has every key capability
    32% less
  • Deepseekv4.1 Flash
    DeepSeek · on Sarvam
    $0.149 / $0.585
    Has every key capability
    33% less

Call it

Using Fireworks AI's API.

Python
from litellm import completion

response = completion(
    model="fireworks_ai/qwen3-235b-a22b-thinking-2507",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Quick answers

How much does Qwen3 235B A22B Thinking 2507 cost?

On Fireworks AI, Qwen3 235B A22B Thinking 2507 costs $0.22 per million input tokens and $0.88 per million output tokens.

Which APIs offer Qwen3 235B A22B Thinking 2507?

Qwen3 235B A22B Thinking 2507 runs on 4 APIs: Fireworks AI, DeepInfra, Novita AI and GMI Cloud.

What is Qwen3 235B A22B Thinking 2507's context window?

262K tokens, with up to 262K tokens in one reply.

Is Qwen3 235B A22B Thinking 2507 being retired?

No retirement date is scheduled for it on any API we track.

What are cheaper alternatives to Qwen3 235B A22B Thinking 2507?

Models that cost less and keep most of what it can do: Claude Haiku 5.5 on Anthropic (48% less), GPT-6 Luna on OpenAI (48% less) and Qwen3.5 397B A17B on AIHubMix (4% less). Savings are on a mix of 3 input tokens to 1 output token.

When was Qwen3 235B A22B Thinking 2507 released?

Alibaba Qwen released it on 25 Jul 2025.

What is Qwen3 235B A22B Thinking 2507's API model id?

Use qwen3-235b-a22b-thinking-2507 with Fireworks AI. Other hosts use their own ids: Qwen3-235B-A22B-Thinking-2507 on DeepInfra, Qwen3-235B-A22B-Thinking-2507-FP8 on GMI Cloud.

Sources

Raw catalog entry for qwen3-235b-a22b-thinking-2507 on Fireworks AI
{
  "max_input_tokens": 262144,
  "max_output_tokens": 262144,
  "input_cost_per_token": 2.2e-7,
  "output_cost_per_token": 8.8e-7,
  "litellm_provider": "fireworks_ai",
  "mode": "chat"
}