Qwen3 capabilities
Qwen3-30B-A3B-FP8 brings Qwen’s hybrid text-generation model into a fine-grained FP8 checkpoint. It accepts text conversations and produces text: explanations, code, translations, creative writing and multi-turn answers. The Qwen3 family supports 119 languages and dialects, including multilingual instructions. Its two response modes let a workflow reserve explicit reasoning for mathematics, programming and logical problems, then use direct answers for straightforward rewriting or dialogue. Language coverage does not imply equal reliability for every language or specialist domain.
Model overview
The model has 30.5 billion parameters in total and activates 3.3 billion for a token. Its mixture-of-experts architecture uses 128 experts, selecting eight per token, across 48 layers. These are model architecture figures; the smaller activated count does not mean that the complete checkpoint contains only 3.3 billion parameters. This is the FP8 edition of Qwen3-30B-A3B, distinct from its original numerical-precision checkpoint and from the later Thinking-2507 and Instruct-2507 releases.
Using the text interface
Prepare the task, relevant source text and expected answer format together. For a debugging request, supply the failing function, observed error, intended behavior and a minimal input, then request an explanation followed by a patch and suggested checks. For multilingual editing, state the target language, audience and terms that must remain unchanged. Ask for a compact final artifact when a long explanation would obscure the result. These are practical ways to use the documented text and coding capabilities, not guarantees that generated code has been executed.
FP8 edition
Qwen specifies fine-grained FP8 quantization with 128-by-128 weight blocks. The checkpoint configuration keeps selected components outside conversion and uses dynamic activation quantization. FP8 describes numerical representation, not a new conversational mode or a promised response time. Compatibility matters: Qwen’s card records a multi-device distributed-inference caveat for this quantization method. A quantized checkpoint should therefore retain its exact edition name rather than silently inherit claims about another build.
Thinking and direct-answer modes
The maker’s chat template enables thinking by default. In that mode, a reasoning block precedes the final answer. A hard enable_thinking=False setting suppresses the block. With thinking enabled, /think and /no_think provide turn-level guidance, following the most recent instruction; an empty reasoning block can still appear with the soft off switch. Those prompt markers cannot override the hard off setting. These are documented maker controls; whether an application exposes them depends on its interface.
Agentic use
Qwen documents tool calling in both modes and recommends Qwen-Agent for the required templates and parsing. Give an agent a clear tool description, permitted inputs and a specific deliverable, such as checking supplied data before drafting a report. The model can participate in a tool workflow; search, code execution and external actions come from the tools connected by the application. Tool capability alone does not give the model access to files or current information.
Long text
Qwen describes a native context of 32,768 tokens and validates extension to 131,072 with YaRN. Input, conversation history, reasoning and answer compete for context space. The shipped position configuration is 40,960; Qwen explains this as an allocation for typical prompts and extended generation, rather than replacing its stated native context. Static YaRN can reduce quality on shorter inputs, so extension is conditional on configuration and need. For document work, label sections and ask the answer to identify the passages supporting each conclusion.
Generation and limitations
Qwen recommends temperature 0.6, top_p 0.95, top_k 20 and min_p 0 for thinking, and 0.7, 0.8, 20 and 0 for direct answers. Greedy decoding can produce degraded reasoning or repeated text. Its guidance recommends a 32,768-token generation budget for most queries and 38,912 for unusually complex problems; these are recommendations, not a separately established hard output ceiling. Larger presence penalties can reduce repetition but may mix languages or lower quality. Keep only final answers in conversation history. State the required answer schema explicitly, and check mathematical results, factual conclusions and generated patches against the supplied evidence before using them.
", "aboutText": "Qwen3 capabilities\n\nQwen3-30B-A3B-FP8 brings Qwen’s hybrid text-generation model into a fine-grained FP8 checkpoint. It accepts text conversations and produces text: explanations, code, translations, creative writing and multi-turn answers. The Qwen3 family supports 119 languages and dialects, including multilingual instructions. Its two response modes let a workflow reserve explicit reasoning for mathematics, programming and logical problems, then use direct answers for straightforward rewriting or dialogue. Language coverage does not imply equal reliability for every language or specialist domain.\n\nModel overview\n\nThe model has 30.5 billion parameters in total and activates 3.3 billion for a token. Its mixture-of-experts architecture uses 128 experts, selecting eight per token, across 48 layers. These are model architecture figures; the smaller activated count does not mean that the complete checkpoint contains only 3.3 billion parameters. This is the FP8 edition of Qwen3-30B-A3B, distinct from its original numerical-precision checkpoint and from the later Thinking-2507 and Instruct-2507 releases.\n\nUsing the text interface\n\nPrepare the task, relevant source text and expected answer format together. For a debugging request, supply the failing function, observed error, intended behavior and a minimal input, then request an explanation followed by a patch and suggested checks. For multilingual editing, state the target language, audience and terms that must remain unchanged. Ask for a compact final artifact when a long explanation would obscure the result. These are practical ways to use the documented text and coding capabilities, not guarantees that generated code has been executed.\n\nFP8 edition\n\nQwen specifies fine-grained FP8 quantization with 128-by-128 weight blocks. The checkpoint configuration keeps selected components outside conversion and uses dynamic activation quantization. FP8 describes numerical representation, not a new conversational mode or a promised response time. Compatibility matters: Qwen’s card records a multi-device distributed-inference caveat for this quantization method. A quantized checkpoint should therefore retain its exact edition name rather than silently inherit claims about another build.\n\nThinking and direct-answer modes\n\nThe maker’s chat template enables thinking by default. In that mode, a reasoning block precedes the final answer. A hard enable_thinking=False setting suppresses the block. With thinking enabled, /think and /no_think provide turn-level guidance, following the most recent instruction; an empty reasoning block can still appear with the soft off switch. Those prompt markers cannot override the hard off setting. These are documented maker controls; whether an application exposes them depends on its interface.\n\nAgentic use\n\nQwen documents tool calling in both modes and recommends Qwen-Agent for the required templates and parsing. Give an agent a clear tool description, permitted inputs and a specific deliverable, such as checking supplied data before drafting a report. The model can participate in a tool workflow; search, code execution and external actions come from the tools connected by the application. Tool capability alone does not give the model access to files or current information.\n\nLong text\n\nQwen describes a native context of 32,768 tokens and validates extension to 131,072 with YaRN. Input, conversation history, reasoning and answer compete for context space. The shipped position configuration is 40,960; Qwen explains this as an allocation for typical prompts and extended generation, rather than replacing its stated native context. Static YaRN can reduce quality on shorter inputs, so extension is conditional on configuration and need. For document work, label sections and ask the answer to identify the passages supporting each conclusion.\n\nGeneration and limitations\n\nQwen recommends temperature 0.6, top_p 0.95, top_k 20 and min_p 0 for thinking, and 0.7, 0.8, 20 and 0 for direct answers. Greedy decoding can produce degraded reasoning or repeated text. Its guidance recommends a 32,768-token generation budget for most queries and 38,912 for unusually complex problems; these are recommendations, not a separately established hard output ceiling. Larger presence penalties can reduce repetition but may mix languages or lower quality. Keep only final answers in conversation history. State the required answer schema explicitly, and check mathematical results, factual conclusions and generated patches against the supplied evidence before using them.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Edit a multilingual service message\n\nTranslate into German for an office administrator: “The Atlas Link update is available, but installing it will restart the device. Save your work before starting.” Keep Atlas Link unchanged, preserve the restart warning and use a polite, concise tone. Return only the translated notice without adding a duration or maintenance window.A concise German notice retaining the product name, restart warning and save-work instruction.Check that the translation keeps the restart as a consequence of installation and does not add an unprovided downtime estimate.qwen/qwen3-30b-a3b-fp8Qwen3-30B-A3B-FP8 is Qwen’s FP8-quantized mixture-of-experts text model, with selectable thinking and direct-answer modes. Use it for mathematical reasoning, coding, multilingual writing and tool-assisted workflows, while preserving the exact FP8 edition and its context configuration.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Qwen3 capabilities
Qwen3-30B-A3B-FP8 brings Qwen’s hybrid text-generation model into a fine-grained FP8 checkpoint. It accepts text conversations and produces text: explanations, code, translations, creative writing and multi-turn answers. The Qwen3 family supports 119 languages and dialects, including multilingual instructions. Its two response modes let a workflow reserve explicit reasoning for mathematics, programming and logical problems, then use direct answers for straightforward rewriting or dialogue. Language coverage does not imply equal reliability for every language or specialist domain.
Model overview
The model has 30.5 billion parameters in total and activates 3.3 billion for a token. Its mixture-of-experts architecture uses 128 experts, selecting eight per token, across 48 layers. These are model architecture figures; the smaller activated count does not mean that the complete checkpoint contains only 3.3 billion parameters. This is the FP8 edition of Qwen3-30B-A3B, distinct from its original numerical-precision checkpoint and from the later Thinking-2507 and Instruct-2507 releases.
Using the text interface
Prepare the task, relevant source text and expected answer format together. For a debugging request, supply the failing function, observed error, intended behavior and a minimal input, then request an explanation followed by a patch and suggested checks. For multilingual editing, state the target language, audience and terms that must remain unchanged. Ask for a compact final artifact when a long explanation would obscure the result. These are practical ways to use the documented text and coding capabilities, not guarantees that generated code has been executed.
FP8 edition
Qwen specifies fine-grained FP8 quantization with 128-by-128 weight blocks. The checkpoint configuration keeps selected components outside conversion and uses dynamic activation quantization. FP8 describes numerical representation, not a new conversational mode or a promised response time. Compatibility matters: Qwen’s card records a multi-device distributed-inference caveat for this quantization method. A quantized checkpoint should therefore retain its exact edition name rather than silently inherit claims about another build.
Thinking and direct-answer modes
The maker’s chat template enables thinking by default. In that mode, a reasoning block precedes the final answer. A hard enable_thinking=False setting suppresses the block. With thinking enabled, /think and /no_think provide turn-level guidance, following the most recent instruction; an empty reasoning block can still appear with the soft off switch. Those prompt markers cannot override the hard off setting. These are documented maker controls; whether an application exposes them depends on its interface.
Agentic use
Qwen documents tool calling in both modes and recommends Qwen-Agent for the required templates and parsing. Give an agent a clear tool description, permitted inputs and a specific deliverable, such as checking supplied data before drafting a report. The model can participate in a tool workflow; search, code execution and external actions come from the tools connected by the application. Tool capability alone does not give the model access to files or current information.
Long text
Qwen describes a native context of 32,768 tokens and validates extension to 131,072 with YaRN. Input, conversation history, reasoning and answer compete for context space. The shipped position configuration is 40,960; Qwen explains this as an allocation for typical prompts and extended generation, rather than replacing its stated native context. Static YaRN can reduce quality on shorter inputs, so extension is conditional on configuration and need. For document work, label sections and ask the answer to identify the passages supporting each conclusion.
Generation and limitations
Qwen recommends temperature 0.6, top_p 0.95, top_k 20 and min_p 0 for thinking, and 0.7, 0.8, 20 and 0 for direct answers. Greedy decoding can produce degraded reasoning or repeated text. Its guidance recommends a 32,768-token generation budget for most queries and 38,912 for unusually complex problems; these are recommendations, not a separately established hard output ceiling. Larger presence penalties can reduce repetition but may mix languages or lower quality. Keep only final answers in conversation history. State the required answer schema explicitly, and check mathematical results, factual conclusions and generated patches against the supplied evidence before using them.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3-30b-a3b-fp8 | Default— |
Fieldmessages | Typemessage[] | RequirementRequired | Values & limits Minimum items: 1 | Default— |
Fieldmessages[].role | Typestring | RequirementRequired | Values & limits Examples: system, user, assistant, tool, function Conditions: These roles are explicitly represented by Source compatibility types; the exact model role set remains model-specific. Model-specific | Default— |
Fieldmessages[].content | Typestring|content_part[]|null | RequirementOptional | Values & limits | Default— |
Fieldmessages[].tool_calls | Typetool_call[] | RequirementOptional | Values & limits | Default— |
Fieldmessages[].tool_call_id | Typestring | RequirementOptional | Values & limits Applies when: role=tool; use actual returned ID | Default— |
Fieldmessages[].name | Typestring | RequirementOptional | Values & limits | Default— |
Fieldmessages[].reasoning_content | Typestring | RequirementOptional | Values & limits Applies when: Round-trip actual returned reasoning where this model uses it | Default— |
Fieldmessages[].reasoning_details | Typeobject|array | RequirementOptional | Values & limits Model-specific | Default— |
Fieldmax_tokens | Typeinteger | RequirementOptional | Values & limits Minimum: 1 Model-specific | Default— |
Fieldmax_completion_tokens | Typeinteger | RequirementOptional | Values & limits Minimum: 1 Model-specific | Default— |
Fieldstream | Typeboolean | RequirementOptional | Values & limits | Defaultfalse |
Fieldstream_options.include_usage | Typeboolean | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtemperature | Typenumber | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtop_p | Typenumber | RequirementOptional | Values & limits Model-specific | Default— |
Fieldstop | Typestring|string[] | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtools | Typefunction_tool[] | RequirementOptional | Values & limits Model-specific | Default— |
Fieldtools[].type | Typestring | RequirementOptional | Values & limits function Model-specific | Default— |
Fieldtools[].function.name | Typestring | RequirementRequired | Values & limits Applies when: function tool supplied | Default— |
Fieldtools[].function.description | Typestring | RequirementOptional | Values & limits | Default— |
Fieldtools[].function.parameters | TypeJSON_schema | RequirementOptional | Values & limits | Default— |
Fieldtool_choice | Typestring|object | RequirementOptional | Values & limits Model-specific | Default— |
Fieldparallel_tool_calls | Typeboolean | RequirementOptional | Values & limits Model-specific | Default— |
Fieldresponse_format | Typeobject | RequirementOptional | Values & limits Model-specific | Default— |
Fieldreasoning_effort | Typestring | RequirementOptional | Values & limits Model-specific | Default— |
Fieldreasoning | Typeobject | RequirementOptional | Values & limits Model-specific | Default— |
Fieldservice_tier | Typestring | RequirementOptional | Values & limits Model-specific | Default— |
Fieldinstructions | Typestring | RequirementOptional | Values & limits Compatibility field | Default— |
Fieldfunctions | Typefunction[] | RequirementOptional | Values & limits Model-specific | Default— |
Fieldfunction_call | Typestring|object | RequirementOptional | Values & limits Model-specific | Default— |
Fieldprompt_cache_options | Typeobject | RequirementOptional | Values & limits Model-specific | Default— |
Fieldmodalities | Typestring[] | RequirementOptional | Values & limits text | Default— |
Fieldprovider | Typeobject | RequirementOptional | Values & limits Conditions: provider.options is rejected on this Ultra text route | Default— |
Fieldcache_control.ttl | Typestring | RequirementOptional | Values & limits Conditions: 1h is rejected in actual prompt-cache locations | Default— |
Text · UTF8 text
messages[].content
| Parameter | Contract |
|---|---|
Parametercontent_type | Contractapplication/json |
Parameterrequired | Contractmodel, messages |
Parametermaterial_types | Contracttext |
Parametermessages | Contractmessage array with role/content shape |
Parameterfile_material | Contractactual UTF-8 document inserted as literal text; no file ID inferred |
Parameteroptional_controls | Contractstream, temperature, top_p, stop, max_tokens, max_completion_tokens, reasoning, response_format, tools, tool_choice |
Parametercontrol_scope | Contractmodel-specific; user function tools differ from hosted tools |
Parameteroutput_modalities | Contracttext |
Parameterhosted_operations | Contractfalse |
Parametermultimodal_input_binding | Contractnull |
Lifecycle: synchronous_or_streamed
Send exact model ID with messages.
Context length is not an output cap.
A text file may supply literal text; image/audio/file-object inputs need exact binding.
Tariff references do not add request options. Supported parameters are listed for the selected method.
| Configuration | Rate schedule | Charge |
|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Charge Input · Input tokens $0.0002 / 1K tokens Output · Output tokens $0.001 / 1K tokens Input · Cache read $0 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\npython3 - <<'PY_REQUEST'\nimport json,os\nfrom pathlib import Path\np=json.loads('{"model": "qwen/qwen3-30b-a3b-fp8", "messages": [{"role": "user", "content": "Explain the difference between an observation and a hypothesis in three sentences."}]}')\n\nPath("request.json").write_text(json.dumps(p,ensure_ascii=False))\nPY_REQUEST\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/chat/completions \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -H 'Content-Type: application/json' \\\n --data-binary @request.json -D response-headers.txt -o result.json\npython3 - <<'PY_RESULT'\nimport json\nfrom pathlib import Path\nr=json.loads(Path("result.json").read_text())\nfor x in r.get("choices",[]):\n if x.get("message",{}).get("content"): print(x["message"]["content"])\nPY_RESULT\nfrom openai import OpenAI\nimport os\n\nif not os.environ.get("AILOOT_API_KEY", "").strip(): raise SystemExit("Set AILOOT_API_KEY before making a request")\nclient=OpenAI(base_url="https://api.ailoot.ai/v1",api_key=os.environ["AILOOT_API_KEY"],max_retries=0)\nresult=client.chat.completions.create(model='qwen/qwen3-30b-a3b-fp8',messages=[{"role":"user","content":"Explain observations and hypotheses in three sentences."}])\nimport base64, json\nfrom pathlib import Path\n\ndef save_result(data, stem="result"):\n # Keep every field, including usage, IDs, reasoning and tool requests.\n Path(stem + ".json").write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")\n print(json.dumps(data, ensure_ascii=False, indent=2))\n def save_blob(encoded, suffix, label):\n raw = base64.b64decode(encoded, altchars=b"-_", validate=True)\n ext = "png" if raw.startswith(b"\\x89PNG\\r\\n\\x1a\\n") else "jpg" if raw.startswith(b"\\xff\\xd8") else "webp" if raw[:4] == b"RIFF" and raw[8:12] == b"WEBP" else suffix\n file = Path(stem + "-" + label + "." + ext)\n file.write_bytes(raw)\n print("Saved", file)\n for index, block in enumerate(data.get("content") or []):\n print("Content", index, json.dumps(block, ensure_ascii=False))\n for choice in data.get("choices") or []:\n print("Choice", choice.get("index"), json.dumps(choice.get("message", choice.get("delta", {})), ensure_ascii=False))\n for index, item in enumerate(data.get("output") or []):\n print("Output", index, json.dumps(item, ensure_ascii=False))\n if item.get("type") == "image_generation_call" and item.get("result"):\n save_blob(item["result"], item.get("output_format") or "bin", "output-" + str(index))\n for ci, candidate in enumerate(data.get("candidates") or []):\n for pi, part in enumerate((candidate.get("content") or {}).get("parts") or []):\n print("Candidate/part", ci, pi, json.dumps(part, ensure_ascii=False))\n blob = part.get("inlineData") or part.get("inline_data")\n if blob and blob.get("data"):\n mime = blob.get("mimeType") or blob.get("mime_type")\n suffix = {"image/png":"png", "image/jpeg":"jpg", "image/webp":"webp", "audio/wav":"wav", "audio/mpeg":"mp3", "video/mp4":"mp4"}.get(mime, "bin")\n save_blob(blob["data"], suffix, "candidate-" + str(ci) + "-part-" + str(pi))\n\nsave_result(result.to_dict(mode="json"))\nfrom openai import OpenAI\nimport os\nfrom pathlib import Path\n\nif not os.environ.get("AILOOT_API_KEY", "").strip(): raise SystemExit("Set AILOOT_API_KEY before making a request")\nclient=OpenAI(base_url="https://api.ailoot.ai/v1",api_key=os.environ["AILOOT_API_KEY"],max_retries=0)\nimport base64, json\nfrom pathlib import Path\n\ndef save_result(data, stem="result"):\n # Keep every field, including usage, IDs, reasoning and tool requests.\n Path(stem + ".json").write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")\n print(json.dumps(data, ensure_ascii=False, indent=2))\n def save_blob(encoded, suffix, label):\n raw = base64.b64decode(encoded, altchars=b"-_", validate=True)\n ext = "png" if raw.startswith(b"\\x89PNG\\r\\n\\x1a\\n") else "jpg" if raw.startswith(b"\\xff\\xd8") else "webp" if raw[:4] == b"RIFF" and raw[8:12] == b"WEBP" else suffix\n file = Path(stem + "-" + label + "." + ext)\n file.write_bytes(raw)\n print("Saved", file)\n for index, block in enumerate(data.get("content") or []):\n print("Content", index, json.dumps(block, ensure_ascii=False))\n for choice in data.get("choices") or []:\n print("Choice", choice.get("index"), json.dumps(choice.get("message", choice.get("delta", {})), ensure_ascii=False))\n for index, item in enumerate(data.get("output") or []):\n print("Output", index, json.dumps(item, ensure_ascii=False))\n if item.get("type") == "image_generation_call" and item.get("result"):\n save_blob(item["result"], item.get("output_format") or "bin", "output-" + str(index))\n for ci, candidate in enumerate(data.get("candidates") or []):\n for pi, part in enumerate((candidate.get("content") or {}).get("parts") or []):\n print("Candidate/part", ci, pi, json.dumps(part, ensure_ascii=False))\n blob = part.get("inlineData") or part.get("inline_data")\n if blob and blob.get("data"):\n mime = blob.get("mimeType") or blob.get("mime_type")\n suffix = {"image/png":"png", "image/jpeg":"jpg", "image/webp":"webp", "audio/wav":"wav", "audio/mpeg":"mp3", "video/mp4":"mp4"}.get(mime, "bin")\n save_blob(blob["data"], suffix, "candidate-" + str(ci) + "-part-" + str(pi))\n\nwith client.chat.completions.stream(model='qwen/qwen3-30b-a3b-fp8',messages=[{"role":"user","content":"Explain observations and hypotheses in three sentences."}]) as stream:\n with Path("stream-events.jsonl").open("w",encoding="utf-8") as log:\n for event in stream:\n payload=event.to_dict(mode="json")\n log.write(json.dumps(payload,ensure_ascii=False)+"\\n"); log.flush()\n print(json.dumps(payload,ensure_ascii=False))\n final=stream.get_final_completion()\n save_result(final.to_dict(mode="json"),"stream-result")\njson or SSE
Keep every complete returned response, file and content part.
Keep every choice and its content/tool calls/reasoning. In SSE assemble deltas by choice/tool index, retain the final finish reason and usage.
Keep request/response IDs and headers. No generic text job retrieval URL is supplied.
| Field | Type |
|---|---|
Fieldchoices[].index | Typeinteger |
Fieldchoices[].message.content | Typestring|null |
Fieldchoices[].message.tool_calls | Typetool_call[] |
Fieldchoices[].message.reasoning_content | Typestring |
Fieldchoices[].message.reasoning_details | Typeobject|array |
Fieldchoices[].finish_reason | Typestring|null |
Fieldusage | Typeobject |
| Response | Condition | Next step |
|---|---|---|
Response400 · invalid_request_error | ConditionInvalid required fields, material representation or parameter value | Next stepCorrect the indicated fields before a new creation request. |
Ailoot Ultra qwen/qwen3-30b-a3b-fp8
Use the method paths shown above.
qwen/qwen3-30b-a3b-fp8Qwen3-30B-A3B-FP8 is Qwen’s FP8-quantized mixture-of-experts text model, with selectable thinking and direct-answer modes. Use it for mathematical reasoning, coding, multilingual writing and tool-assisted workflows, while preserving the exact FP8 edition and its context configuration.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Qwen3 capabilities
Qwen3-30B-A3B-FP8 brings Qwen’s hybrid text-generation model into a fine-grained FP8 checkpoint. It accepts text conversations and produces text: explanations, code, translations, creative writing and multi-turn answers. The Qwen3 family supports 119 languages and dialects, including multilingual instructions. Its two response modes let a workflow reserve explicit reasoning for mathematics, programming and logical problems, then use direct answers for straightforward rewriting or dialogue. Language coverage does not imply equal reliability for every language or specialist domain.
Model overview
The model has 30.5 billion parameters in total and activates 3.3 billion for a token. Its mixture-of-experts architecture uses 128 experts, selecting eight per token, across 48 layers. These are model architecture figures; the smaller activated count does not mean that the complete checkpoint contains only 3.3 billion parameters. This is the FP8 edition of Qwen3-30B-A3B, distinct from its original numerical-precision checkpoint and from the later Thinking-2507 and Instruct-2507 releases.
Using the text interface
Prepare the task, relevant source text and expected answer format together. For a debugging request, supply the failing function, observed error, intended behavior and a minimal input, then request an explanation followed by a patch and suggested checks. For multilingual editing, state the target language, audience and terms that must remain unchanged. Ask for a compact final artifact when a long explanation would obscure the result. These are practical ways to use the documented text and coding capabilities, not guarantees that generated code has been executed.
FP8 edition
Qwen specifies fine-grained FP8 quantization with 128-by-128 weight blocks. The checkpoint configuration keeps selected components outside conversion and uses dynamic activation quantization. FP8 describes numerical representation, not a new conversational mode or a promised response time. Compatibility matters: Qwen’s card records a multi-device distributed-inference caveat for this quantization method. A quantized checkpoint should therefore retain its exact edition name rather than silently inherit claims about another build.
Thinking and direct-answer modes
The maker’s chat template enables thinking by default. In that mode, a reasoning block precedes the final answer. A hard enable_thinking=False setting suppresses the block. With thinking enabled, /think and /no_think provide turn-level guidance, following the most recent instruction; an empty reasoning block can still appear with the soft off switch. Those prompt markers cannot override the hard off setting. These are documented maker controls; whether an application exposes them depends on its interface.
Agentic use
Qwen documents tool calling in both modes and recommends Qwen-Agent for the required templates and parsing. Give an agent a clear tool description, permitted inputs and a specific deliverable, such as checking supplied data before drafting a report. The model can participate in a tool workflow; search, code execution and external actions come from the tools connected by the application. Tool capability alone does not give the model access to files or current information.
Long text
Qwen describes a native context of 32,768 tokens and validates extension to 131,072 with YaRN. Input, conversation history, reasoning and answer compete for context space. The shipped position configuration is 40,960; Qwen explains this as an allocation for typical prompts and extended generation, rather than replacing its stated native context. Static YaRN can reduce quality on shorter inputs, so extension is conditional on configuration and need. For document work, label sections and ask the answer to identify the passages supporting each conclusion.
Generation and limitations
Qwen recommends temperature 0.6, top_p 0.95, top_k 20 and min_p 0 for thinking, and 0.7, 0.8, 20 and 0 for direct answers. Greedy decoding can produce degraded reasoning or repeated text. Its guidance recommends a 32,768-token generation budget for most queries and 38,912 for unusually complex problems; these are recommendations, not a separately established hard output ceiling. Larger presence penalties can reduce repetition but may mix languages or lower quality. Keep only final answers in conversation history. State the required answer schema explicitly, and check mathematical results, factual conclusions and generated patches against the supplied evidence before using them.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits qwen/qwen3-30b-a3b-fp8 | Default— |
Fieldmessages | Typemessage[] | RequirementRequired | Values & limits | Default— |
Fieldmessages[].role | Typestring | RequirementRequired | Values & limits user, assistant | Default— |
Fieldmessages[].content | Typestring | RequirementRequired | Values & limits | Default— |
Fieldmax_tokens | Typeinteger | RequirementRequired | Values & limits | Default— |
Fieldstream | Typeboolean | RequirementOptional | Values & limits | Defaultfalse |
Text · UTF-8 text
messages[].content string
| Parameter | Contract |
|---|---|
Parametercontent_type | Contractapplication/json |
Parameterrequired | Contractmodel, messages, max_tokens |
Parameteroptional_controls | Contractstream |
Parametermaterial_types | Contracttext |
Parametertext_file_material | ContractRead the actual UTF-8 document and put its text in the indicated input or message content string. |
Lifecycle: synchronous_or_streamed
Tariff references do not add request options. Supported parameters are listed for the selected method.
| Configuration | Rate schedule | Charge |
|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Charge Input · Input tokens $0.0002 / 1K tokens Output · Output tokens $0.001 / 1K tokens Input · Cache read $0 / 1K tokens Input · Cache write $0 / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "qwen/qwen3-30b-a3b-fp8",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\n: "${AILOOT_TEXT_FILE:?Set your actual UTF-8 document}"\nexport AILOOT_TEXT_FILE\npython3 - <<'PY_REQUEST'\nimport json, os\nfrom pathlib import Path\n\nsource_file = Path(os.environ['AILOOT_TEXT_FILE'])\nif not source_file.is_file():\n raise SystemExit('Set AILOOT_TEXT_FILE to your actual UTF-8 document')\ntext = source_file.read_text(encoding='utf-8')\nif not text.strip():\n raise SystemExit('Choose a document containing text')\npayload = {'model': 'qwen/qwen3-30b-a3b-fp8', 'max_tokens': 128, 'messages': [{'role': 'user', 'content': 'Summarize the incident report.'}]}\npayload['messages'][0]['content'] = 'Summarize this material and quote supporting source passages.\\n\\n' + text\nPath('request.json').write_text(json.dumps(payload, ensure_ascii=False), encoding='utf-8')\nPY_REQUEST\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/messages \\\n -H "x-api-key: $AILOOT_API_KEY" -H 'anthropic-version: 2023-06-01' -H 'Content-Type: application/json' \\\n --data-binary @request.json -D response-headers.txt -o result.json\npython3 - <<'PY_RESULT'\nimport json\nfrom pathlib import Path\n\nresult = json.loads(Path('result.json').read_text(encoding='utf-8'))\nprint(json.dumps(result, ensure_ascii=False, indent=2))\nPY_RESULT\nimport json, os\nfrom pathlib import Path\n\nsource_file = Path(os.environ['AILOOT_TEXT_FILE'])\nif not source_file.is_file():\n raise SystemExit('Set AILOOT_TEXT_FILE to your actual UTF-8 document')\ntext = source_file.read_text(encoding='utf-8')\nif not text.strip():\n raise SystemExit('Choose a document containing text')\npayload = {'model': 'qwen/qwen3-30b-a3b-fp8', 'max_tokens': 128, 'messages': [{'role': 'user', 'content': 'Summarize the incident report.'}]}\npayload['messages'][0]['content'] = 'Summarize this material and quote supporting source passages.\\n\\n' + text\nfrom urllib.request import Request, urlopen\nfrom urllib.error import HTTPError\n\napi_key = os.environ['AILOOT_API_KEY']\nif not api_key.strip():\n raise SystemExit('Set AILOOT_API_KEY before making a request')\nheaders = {'x-api-key': 'PLACEHOLDER_KEY', 'anthropic-version': '2023-06-01', 'Content-Type': 'application/json'}\nheaders['x-api-key'] = api_key\nrequest = Request('https://api.ailoot.ai/v1/messages', data=json.dumps(payload, ensure_ascii=False).encode('utf-8'), headers=headers, method='POST')\ntry:\n with urlopen(request, timeout=120) as response:\n raw = response.read()\n Path('response-headers.txt').write_text(str(response.headers), encoding='utf-8')\nexcept HTTPError as error:\n Path('result.json').write_bytes(error.read())\n Path('response-headers.txt').write_text(str(error.headers), encoding='utf-8')\n raise\nPath('result.json').write_bytes(raw)\nimport json\nfrom pathlib import Path\n\nresult = json.loads(Path('result.json').read_text(encoding='utf-8'))\nprint(json.dumps(result, ensure_ascii=False, indent=2))\nJSON or SSE
Keep every complete returned response, file and content part.
Keep all content blocks, thinking signatures and actual tool-use IDs. Assemble all SSE block deltas; a first text block is not the complete output.
Keep request/response IDs and headers. No generic text job retrieval URL is supplied.
| Field | Type |
|---|---|
Fieldid | Typestring |
Fieldmodel | Typestring |
Fieldcontent[] | Typetext|thinking|tool_use|other content block |
Fieldstop_reason | Typestring|null |
Fieldusage | Typeobject |
| Response | Condition | Next step |
|---|---|---|
Response400 · invalid_request_error | ConditionInvalid required fields, material representation or parameter value | Next stepCorrect the indicated fields before a new creation request. |
Ailoot Ultra qwen/qwen3-30b-a3b-fp8
Use the method paths shown above.
Technical example in English
Edit a multilingual service message\n\nTranslate into German for an office administrator: “The Atlas Link update is available, but installing it will restart the device. Save your work before starting.” Keep Atlas Link unchanged, preserve the restart warning and use a polite, concise tone. Return only the translated notice without adding a duration or maintenance window.A concise German notice retaining the product name, restart warning and save-work instruction.Check that the translation keeps the restart as a consequence of installation and does not add an unprovided downtime estimate.