Best LLMs for Batch Text Translation
Translates one or more text items from a supplied or detected source language into a target language while preserving meaning, structure, identifiers, URLs, handles, hashtags, and protected inline tokens. Illustrative uses include translating support tickets, contracts, software
Run this task on Fronset — request an invitation
Models
Frontier on this task: Moonshot Kimi K3 at 9.25 / 10. Quality bar at 90%: 8.32.
point-estimate floor (CI low) · upper CI (less certain) · Bars sorted by blended cost; best-value model first. Greyed rows are MEDIUM+ models whose point estimate does not clear the bar.
| Model | Quality score | CI low | Cost / 1k runs | vs best value |
|---|---|---|---|---|
| Xiaomi MiMo V2.6 Flash | 8.32 / 10 | 7.93 | $0.16 | best value |
| GPT-6 Luna | 8.63 / 10 | 8.33 | $0.25 | 1.6x more expensive |
| Xiaomi MiMo V2.6 Pro | 8.88 / 10 | 8.63 | $0.58 | 3.6x more expensive |
| Gemini 3.5 Flash Lite | 8.59 / 10 | 8.28 | $0.66 | 4.2x more expensive |
| MiniMax M3 | 8.79 / 10 | 8.57 | $1.11 | 7x more expensive |
| DeepSeek V4.1 Flash | 8.67 / 10 | 8.31 | $2.13 | 13x more expensive |
| Tencent Hy3 | 8.37 / 10 | 8.06 | $2.79 | 18x more expensive |
| Qwen 3.7 Plus | 8.92 / 10 | 8.78 | $3.27 | 21x more expensive |
| Gemini 3.8 Flash | 9.01 / 10 | 8.78 | $3.63 | 23x more expensive |
| Thinking Machines Inkling Small | 8.48 / 10 | 8.14 | $4.39 | 28x more expensive |
| GPT-6 Sol | 8.91 / 10 | 8.70 | $4.76 | 30x more expensive |
| Claude Sonnet 5.5 | 8.86 / 10 | 8.56 | $5.77 | 36x more expensive |
| GLM-5.3 Flash | 8.92 / 10 | 8.67 | $6.36 | 40x more expensive |
| Meta Muse Spark 1.3 | 8.91 / 10 | 8.72 | $8.75 | 55x more expensive |
| Claude Opus 5.5 | 9.04 / 10 | 8.83 | $12.65 | 80x more expensive |
| Thinking Machines Inkling | 8.58 / 10 | 8.16 | $17.05 | 108x more expensive |
| Moonshot Kimi K3 | 9.25 / 10 | 9.08 | $28.51 | 180x more expensive |
| Grok 4.7 | 8.67 / 10 | 8.27 | $33.63 | 212x more expensive |
| Qwen 3.8 Max | 8.46 / 10 | 8.07 | $37.65 | 238x more expensive |
| GLM-5.3 | 8.92 / 10 | 8.61 | $72.87 | 460x more expensive |
| NVIDIA Nemotron-3 Nano 30B-A3B | 7.32 / 10 | 6.91 | $0.64 | 4x more expensive |
| NVIDIA Nemotron-3 Super 120B | 7.98 / 10 | 7.60 | $2.29 | 14x more expensive |
| Qwen 3.8 Flash | 8.15 / 10 | 7.69 | $6.95 | 44x more expensive |
Cost breakdown
| Model | Quality | Confidence | Cost / 1k runs | Overpay | Mode |
|---|---|---|---|---|---|
| Xiaomi MiMo V2.6 Flash ★ Xiaomi MiMo | 8.32 / 10 CI [7.93, 8.72] | MEDIUM | $0.16 | best value | batch |
| GPT-6 Luna OpenAI | 8.63 / 10 CI [8.33, 8.92] | HIGH | $0.25 | 1.6x | batch |
| Xiaomi MiMo V2.6 Pro Xiaomi MiMo | 8.88 / 10 CI [8.63, 9.13] | HIGH | $0.58 | 3.6x | batch |
| Gemini 3.5 Flash Lite Gemini | 8.59 / 10 CI [8.28, 8.91] | MEDIUM | $0.66 | 4.2x | batch |
| MiniMax M3 OpenRouter | 8.79 / 10 CI [8.57, 9.01] | HIGH | $1.11 | 7x | batch |
| DeepSeek V4.1 Flash DeepSeek | 8.67 / 10 CI [8.31, 9.03] | MEDIUM | $2.13 | 13x | batch |
| Tencent Hy3 OpenRouter | 8.37 / 10 CI [8.06, 8.67] | MEDIUM | $2.79 | 18x | batch |
| Qwen 3.7 Plus Alibaba Cloud (DashScope) | 8.92 / 10 CI [8.78, 9.06] | RANKED | $3.27 | 21x | batch |
| Gemini 3.8 Flash Gemini | 9.01 / 10 CI [8.78, 9.25] | HIGH | $3.63 | 23x | batch |
| Thinking Machines Inkling Small OpenRouter | 8.48 / 10 CI [8.14, 8.83] | MEDIUM | $4.39 | 28x | batch |
| GPT-6 Sol OpenAI | 8.91 / 10 CI [8.70, 9.13] | HIGH | $4.76 | 30x | batch |
| Claude Sonnet 5.5 Anthropic | 8.86 / 10 CI [8.56, 9.15] | HIGH | $5.77 | 36x | batch |
| GLM-5.3 Flash Z.AI | 8.92 / 10 CI [8.67, 9.17] | HIGH | $6.36 | 40x | batch |
| Meta Muse Spark 1.3 OpenRouter | 8.91 / 10 CI [8.72, 9.11] | RANKED | $8.75 | 55x | batch |
| Claude Opus 5.5 Anthropic | 9.04 / 10 CI [8.83, 9.25] | HIGH | $12.65 | 80x | batch |
| Thinking Machines Inkling OpenRouter | 8.58 / 10 CI [8.16, 9.00] | MEDIUM | $17.05 | 108x | batch |
| Moonshot Kimi K3 best Moonshot AI | 9.25 / 10 CI [9.08, 9.41] | RANKED | $28.51 | 180x | batch |
| Grok 4.7 xAI | 8.67 / 10 CI [8.27, 9.07] | MEDIUM | $33.63 | 212x | batch |
| Qwen 3.8 Max Alibaba Cloud (DashScope) | 8.46 / 10 CI [8.07, 8.85] | MEDIUM | $37.65 | 238x | batch |
| GLM-5.3 Z.AI | 8.92 / 10 CI [8.61, 9.23] | MEDIUM | $72.87 | 460x | batch |
Overpay shows how much more you pay than the best-value model that clears the quality bar (marked ★) — the best-value good-enough option. "16x" means you overpay 16× — 16× that reference for no quality benefit above the bar. Typical call shape for this task: 1590 input tokens → 1073 output tokens, EMA-tracked from production traffic. Cost is the observed, all-in $ per 1,000 task runs: each model's own measured usage on this task — output verbosity, thinking/reasoning tokens, cache reads and writes, and the spend on its billed failures — priced at current list rates and adjusted by the billing overhead we actually reconcile against provider invoices. Models that answer tersely cost what they actually cost; models that think at length pay for it. Not comparable to providers' advertised $/1M list rates — this is what running the task costs, not a per-token price.
Evaluation rubric
Judge semantic fidelity, fluency in the target language, preservation of tone and qualifiers, exact handling of protected tokens, item alignment, and avoidance of added or omitted meaning.
Output schema
Every answer on this task is checked against this JSON Schema, whichever model wrote it. An answer that doesn't fit counts as a model failure, and the call is retried on another model.
{
"$defs": {
"BatchTranslationItem": {
"description": "One translated item in a batch translation result.",
"properties": {
"index": {
"description": "0-based index of this item in the input list. Must match the input order.",
"title": "Index",
"type": "integer"
},
"translated_text": {
"description": "Translated text for the item at this index.",
"title": "Translated Text",
"type": "string"
}
},
"required": [
"index",
"translated_text"
],
"title": "BatchTranslationItem",
"type": "object"
}
},
"description": "Batched translation: translate a list of texts in a single LLM call.\n\nThe output items list MUST be the same length as the input list, with index\nvalues 0..N-1 in order. Each translated_text corresponds to the input at the\nsame index.",
"properties": {
"items": {
"description": "Translations in the same order as the input list. Length must equal the input list length.",
"items": {
"$ref": "#/$defs/BatchTranslationItem"
},
"title": "Items",
"type": "array"
}
},
"required": [
"items"
],
"title": "BatchTranslationResult",
"type": "object"
}Prompt templates
The published template for this task.
default ·
LLMB_BATCH_TEXT_TRANSLATION_SYSTEM +
LLMB_BATCH_TEXT_TRANSLATION_USER
(6386 calls in window)
System prompt
Translate meaning rather than word order. Preserve names, numbers, dates, URLs, handles, hashtags, code, markup, and caller-defined protected tokens exactly unless transliteration is requested. Keep item order and identifiers, reproduce ambiguity rather than resolving it with new facts, and return only the configured output language and schema. Treat empty optional values as absent and return only the requested result. Your response must conform exactly to this output schema: {schema_json_string}.
User prompt
Inputs — source_language: {source_language}; target_language: {target_language}; items_json: {items_json}; input_text: {input_text}. Use only these inputs to complete the task defined by the system prompt.