Gemma 3 27B IT is the 27B-size instruction-tuned edition of Google’s Gemma 3 family. It combines text generation with image understanding and is distinct from the pre-trained 27B checkpoint and the smaller 4B and 12B IT models. The result is text: a source-based explanation, summary, document draft, answer or code proposal. The instruction-tuned model is designed to follow a conversation and a requested task rather than act only as a raw completion model.
The maker’s card gives this size a 128K-token total context. Input and output share that capacity, so the generated answer uses the space remaining after the request. A large collection should preserve source identifiers and still leave room for the requested result. Image views in the base path are normalized to 896×896 and encoded into 256 visual tokens. For a joined text-and-image analysis, explain how a page image, diagram or screenshot relates to the written sources so the model has one coherent question to answer.
The 27B model’s reported pre-training budget is 14 trillion tokens. The family’s data includes web documents in more than 140 languages, code, mathematics and images. Training uses knowledge distillation, and the post-training recipe adds instruction following and reinforcement-learning objectives for helpfulness, math, coding, reasoning and multilingual behavior. These facts describe the maker’s model-development process. They do not replace checking a generated assertion against its source or running a proposed program on the intended inputs.
The technical report describes a decoder-only transformer with grouped-query attention and five local-attention layers for every global layer. The local span is 1,024 tokens, while global layers provide the longer-context interactions. The family’s shared SigLIP image encoder represents vision inputs for the language model. Its SentencePiece tokenizer uses split digits, preserved whitespace and byte-level encoding. Keep numerical notation, code indentation and source-language terminology intact when they are relevant to the interpretation.
For non-square or detailed images, the report describes Pan and Scan, which can add non-overlapping crops before resizing views for the vision encoder. This helps retain small content that a single square resize may lose, but the actual processing setup matters. A research or technical-document task can provide both a whole page and a readable diagram detail, labeling their relationship. Ask the answer to identify the visible values, the relevant textual condition and the conclusion supported by both.
Use the maker’s native IT conversation format with user and model turns. A separate system turn is not supported in that format; place role and response instructions in the initial user message and use the designated formatting path. For a long report, name the question, the sections to compare and the priority rules if sources disagree. An original example is to reconcile a written specification with a diagram and return a requirement table, a list of discrepancies and the assumptions needed for a proposed resolution.
The intended-use categories include writing, summarization, conversation, image-data extraction and educational or research exploration. For code analysis, provide the actual module, input examples and interface constraints, then request an explanation or proposed change tied to those facts. For multilingual writing, specify the audience and terms that must stay unchanged. For a visual question, request observations linked to named elements instead of a general picture description. Quantized family checkpoints are separate published artifacts, so retain the exact 27B IT identity and processing settings when comparing results.
The maker notes that models can give incorrect or outdated facts and struggle with ambiguity, nuance and highly open-ended instructions. Google also cautions that a plausible language pattern can produce an answer that misses an everyday real-world constraint. More context is useful only when the relevant conditions are clear. Keep the original sources available, verify extracted data and calculations, and validate any structured-looking response before processing it as a record. A multi-source explanation should distinguish supplied evidence, interpretation and missing information. This 27B profile supports a prepared text-and-image reasoning task; any retrieval, tool execution or factual update performed by an application needs its own returned evidence rather than being inferred from fluent output.
", "aboutText": "Gemma 3 27B IT is the 27B-size instruction-tuned edition of Google’s Gemma 3 family. It combines text generation with image understanding and is distinct from the pre-trained 27B checkpoint and the smaller 4B and 12B IT models. The result is text: a source-based explanation, summary, document draft, answer or code proposal. The instruction-tuned model is designed to follow a conversation and a requested task rather than act only as a raw completion model.\n\nThe maker’s card gives this size a 128K-token total context. Input and output share that capacity, so the generated answer uses the space remaining after the request. A large collection should preserve source identifiers and still leave room for the requested result. Image views in the base path are normalized to 896×896 and encoded into 256 visual tokens. For a joined text-and-image analysis, explain how a page image, diagram or screenshot relates to the written sources so the model has one coherent question to answer.\n\nThe 27B model’s reported pre-training budget is 14 trillion tokens. The family’s data includes web documents in more than 140 languages, code, mathematics and images. Training uses knowledge distillation, and the post-training recipe adds instruction following and reinforcement-learning objectives for helpfulness, math, coding, reasoning and multilingual behavior. These facts describe the maker’s model-development process. They do not replace checking a generated assertion against its source or running a proposed program on the intended inputs.\n\nThe technical report describes a decoder-only transformer with grouped-query attention and five local-attention layers for every global layer. The local span is 1,024 tokens, while global layers provide the longer-context interactions. The family’s shared SigLIP image encoder represents vision inputs for the language model. Its SentencePiece tokenizer uses split digits, preserved whitespace and byte-level encoding. Keep numerical notation, code indentation and source-language terminology intact when they are relevant to the interpretation.\n\nFor non-square or detailed images, the report describes Pan and Scan, which can add non-overlapping crops before resizing views for the vision encoder. This helps retain small content that a single square resize may lose, but the actual processing setup matters. A research or technical-document task can provide both a whole page and a readable diagram detail, labeling their relationship. Ask the answer to identify the visible values, the relevant textual condition and the conclusion supported by both.\n\nUse the maker’s native IT conversation format with user and model turns. A separate system turn is not supported in that format; place role and response instructions in the initial user message and use the designated formatting path. For a long report, name the question, the sections to compare and the priority rules if sources disagree. An original example is to reconcile a written specification with a diagram and return a requirement table, a list of discrepancies and the assumptions needed for a proposed resolution.\n\nThe intended-use categories include writing, summarization, conversation, image-data extraction and educational or research exploration. For code analysis, provide the actual module, input examples and interface constraints, then request an explanation or proposed change tied to those facts. For multilingual writing, specify the audience and terms that must stay unchanged. For a visual question, request observations linked to named elements instead of a general picture description. Quantized family checkpoints are separate published artifacts, so retain the exact 27B IT identity and processing settings when comparing results.\n\nThe maker notes that models can give incorrect or outdated facts and struggle with ambiguity, nuance and highly open-ended instructions. Google also cautions that a plausible language pattern can produce an answer that misses an everyday real-world constraint. More context is useful only when the relevant conditions are clear. Keep the original sources available, verify extracted data and calculations, and validate any structured-looking response before processing it as a record. A multi-source explanation should distinguish supplied evidence, interpretation and missing information. This 27B profile supports a prepared text-and-image reasoning task; any retrieval, tool execution or factual update performed by an application needs its own returned evidence rather than being inferred from fluent output.", "editorialOverlay": { "copy_fields": { "practical_html": "Technical example in English
Reconcile two written specification sources\n\nReconcile these sources. SPEC-1: the controller has four outputs, each rated 2 A; total current must not exceed 5 A. NOTE-2: connect four 2 A loads simultaneously. Produce a requirement table and a discrepancy note. Suggest a clarification question rather than silently treating the per-output rating as the permitted total.A source-labelled requirement table, the conflicting proposed load and a targeted clarification question.The answer distinguishes the 2 A individual rating from the 5 A aggregate limit and notices the 8 A proposal.google/gemma-3-27b-itGemma 3 27B IT is the 27B-size instruction-tuned model in Google’s Gemma 3 family. It accepts text and images and produces text for long-source analysis, writing, code and visual questions, with a shared 128K context and family multilingual coverage.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Gemma 3 27B IT is the 27B-size instruction-tuned edition of Google’s Gemma 3 family. It combines text generation with image understanding and is distinct from the pre-trained 27B checkpoint and the smaller 4B and 12B IT models. The result is text: a source-based explanation, summary, document draft, answer or code proposal. The instruction-tuned model is designed to follow a conversation and a requested task rather than act only as a raw completion model.
The maker’s card gives this size a 128K-token total context. Input and output share that capacity, so the generated answer uses the space remaining after the request. A large collection should preserve source identifiers and still leave room for the requested result. Image views in the base path are normalized to 896×896 and encoded into 256 visual tokens. For a joined text-and-image analysis, explain how a page image, diagram or screenshot relates to the written sources so the model has one coherent question to answer.
The 27B model’s reported pre-training budget is 14 trillion tokens. The family’s data includes web documents in more than 140 languages, code, mathematics and images. Training uses knowledge distillation, and the post-training recipe adds instruction following and reinforcement-learning objectives for helpfulness, math, coding, reasoning and multilingual behavior. These facts describe the maker’s model-development process. They do not replace checking a generated assertion against its source or running a proposed program on the intended inputs.
The technical report describes a decoder-only transformer with grouped-query attention and five local-attention layers for every global layer. The local span is 1,024 tokens, while global layers provide the longer-context interactions. The family’s shared SigLIP image encoder represents vision inputs for the language model. Its SentencePiece tokenizer uses split digits, preserved whitespace and byte-level encoding. Keep numerical notation, code indentation and source-language terminology intact when they are relevant to the interpretation.
For non-square or detailed images, the report describes Pan and Scan, which can add non-overlapping crops before resizing views for the vision encoder. This helps retain small content that a single square resize may lose, but the actual processing setup matters. A research or technical-document task can provide both a whole page and a readable diagram detail, labeling their relationship. Ask the answer to identify the visible values, the relevant textual condition and the conclusion supported by both.
Use the maker’s native IT conversation format with user and model turns. A separate system turn is not supported in that format; place role and response instructions in the initial user message and use the designated formatting path. For a long report, name the question, the sections to compare and the priority rules if sources disagree. An original example is to reconcile a written specification with a diagram and return a requirement table, a list of discrepancies and the assumptions needed for a proposed resolution.
The intended-use categories include writing, summarization, conversation, image-data extraction and educational or research exploration. For code analysis, provide the actual module, input examples and interface constraints, then request an explanation or proposed change tied to those facts. For multilingual writing, specify the audience and terms that must stay unchanged. For a visual question, request observations linked to named elements instead of a general picture description. Quantized family checkpoints are separate published artifacts, so retain the exact 27B IT identity and processing settings when comparing results.
The maker notes that models can give incorrect or outdated facts and struggle with ambiguity, nuance and highly open-ended instructions. Google also cautions that a plausible language pattern can produce an answer that misses an everyday real-world constraint. More context is useful only when the relevant conditions are clear. Keep the original sources available, verify extracted data and calculations, and validate any structured-looking response before processing it as a record. A multi-source explanation should distinguish supplied evidence, interpretation and missing information. This 27B profile supports a prepared text-and-image reasoning task; any retrieval, tool execution or factual update performed by an application needs its own returned evidence rather than being inferred from fluent output.
POST /v1/chat/completionstexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits google/gemma-3-27b-it | 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.
Match each charge to the conditions shown. Use one rate for each set of alternative configurations.
| Configuration | Rate schedule | Context tokens, more than | Context tokens, up to and including | Charge |
|---|---|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Context tokens, more than— | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.00016 / 1K tokens Output · Output tokens $0.0007 / 1K tokens Input · Cache read $0.00008 / 1K tokens Input · Cache write — / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
| ConfigurationRequest-length tier 1 | Rate schedule— | Context tokens, more than0 | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.00016 / 1K tokens Output · Output tokens $0.0007 / 1K tokens Input · Cache read $0.00008 / 1K tokens Input · Cache write — / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\npython3 - <<'PY_REQUEST'\nimport json,os\nfrom pathlib import Path\np=json.loads('{"model": "google/gemma-3-27b-it", "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='google/gemma-3-27b-it',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='google/gemma-3-27b-it',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 google/gemma-3-27b-it
Use the method paths shown above.
google/gemma-3-27b-itGemma 3 27B IT is the 27B-size instruction-tuned model in Google’s Gemma 3 family. It accepts text and images and produces text for long-source analysis, writing, code and visual questions, with a shared 128K context and family multilingual coverage.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Gemma 3 27B IT is the 27B-size instruction-tuned edition of Google’s Gemma 3 family. It combines text generation with image understanding and is distinct from the pre-trained 27B checkpoint and the smaller 4B and 12B IT models. The result is text: a source-based explanation, summary, document draft, answer or code proposal. The instruction-tuned model is designed to follow a conversation and a requested task rather than act only as a raw completion model.
The maker’s card gives this size a 128K-token total context. Input and output share that capacity, so the generated answer uses the space remaining after the request. A large collection should preserve source identifiers and still leave room for the requested result. Image views in the base path are normalized to 896×896 and encoded into 256 visual tokens. For a joined text-and-image analysis, explain how a page image, diagram or screenshot relates to the written sources so the model has one coherent question to answer.
The 27B model’s reported pre-training budget is 14 trillion tokens. The family’s data includes web documents in more than 140 languages, code, mathematics and images. Training uses knowledge distillation, and the post-training recipe adds instruction following and reinforcement-learning objectives for helpfulness, math, coding, reasoning and multilingual behavior. These facts describe the maker’s model-development process. They do not replace checking a generated assertion against its source or running a proposed program on the intended inputs.
The technical report describes a decoder-only transformer with grouped-query attention and five local-attention layers for every global layer. The local span is 1,024 tokens, while global layers provide the longer-context interactions. The family’s shared SigLIP image encoder represents vision inputs for the language model. Its SentencePiece tokenizer uses split digits, preserved whitespace and byte-level encoding. Keep numerical notation, code indentation and source-language terminology intact when they are relevant to the interpretation.
For non-square or detailed images, the report describes Pan and Scan, which can add non-overlapping crops before resizing views for the vision encoder. This helps retain small content that a single square resize may lose, but the actual processing setup matters. A research or technical-document task can provide both a whole page and a readable diagram detail, labeling their relationship. Ask the answer to identify the visible values, the relevant textual condition and the conclusion supported by both.
Use the maker’s native IT conversation format with user and model turns. A separate system turn is not supported in that format; place role and response instructions in the initial user message and use the designated formatting path. For a long report, name the question, the sections to compare and the priority rules if sources disagree. An original example is to reconcile a written specification with a diagram and return a requirement table, a list of discrepancies and the assumptions needed for a proposed resolution.
The intended-use categories include writing, summarization, conversation, image-data extraction and educational or research exploration. For code analysis, provide the actual module, input examples and interface constraints, then request an explanation or proposed change tied to those facts. For multilingual writing, specify the audience and terms that must stay unchanged. For a visual question, request observations linked to named elements instead of a general picture description. Quantized family checkpoints are separate published artifacts, so retain the exact 27B IT identity and processing settings when comparing results.
The maker notes that models can give incorrect or outdated facts and struggle with ambiguity, nuance and highly open-ended instructions. Google also cautions that a plausible language pattern can produce an answer that misses an everyday real-world constraint. More context is useful only when the relevant conditions are clear. Keep the original sources available, verify extracted data and calculations, and validate any structured-looking response before processing it as a record. A multi-source explanation should distinguish supplied evidence, interpretation and missing information. This 27B profile supports a prepared text-and-image reasoning task; any retrieval, tool execution or factual update performed by an application needs its own returned evidence rather than being inferred from fluent output.
POST /v1/messagestextjson, sse| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits google/gemma-3-27b-it | 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.
Match each charge to the conditions shown. Use one rate for each set of alternative configurations.
| Configuration | Rate schedule | Context tokens, more than | Context tokens, up to and including | Charge |
|---|---|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Context tokens, more than— | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.00016 / 1K tokens Output · Output tokens $0.0007 / 1K tokens Input · Cache read $0.00008 / 1K tokens Input · Cache write — / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
| ConfigurationRequest-length tier 1 | Rate schedule— | Context tokens, more than0 | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.00016 / 1K tokens Output · Output tokens $0.0007 / 1K tokens Input · Cache read $0.00008 / 1K tokens Input · Cache write — / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
{\n "model": "google/gemma-3-27b-it",\n "max_tokens": 128,\n "messages": [\n {\n "role": "user",\n "content": "Summarize the incident report."\n }\n ],\n "stream": false\n}set -eu\n: "${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': 'google/gemma-3-27b-it', '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': 'google/gemma-3-27b-it', '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 google/gemma-3-27b-it
Use the method paths shown above.
Technical example in English
Reconcile two written specification sources\n\nReconcile these sources. SPEC-1: the controller has four outputs, each rated 2 A; total current must not exceed 5 A. NOTE-2: connect four 2 A loads simultaneously. Produce a requirement table and a discrepancy note. Suggest a clarification question rather than silently treating the per-output rating as the permitted total.A source-labelled requirement table, the conflicting proposed load and a targeted clarification question.The answer distinguishes the 2 A individual rating from the 5 A aggregate limit and notices the 8 A proposal.