Best LLM for Translation in 2026
For translation, Muse Spark 1.3 Contributor is our pick: $0.15/M tokens on a Translation batch workload, 1.0M context.
Translation workloads are typically high-volume and roughly balanced between input and output tokens, so we weight raw price the heaviest here, with speed as the next-biggest lever for real-time use cases.
Quick answer: What is the best LLM for translation?
Muse Spark 1.3 Contributor, from Meta, is the best fit for translation at $0.15 per million task tokens on a Translation batch workload, with a 1.0M-token context window. No cheaper value pick beats it for this task.
What evidence supports the Translation recommendation?
Reproducible Translation evidence and decision rubric
| Test / run | Prompt and verification | Hard rule |
|---|---|---|
| Code Snippet | exact prompt + 20 recorded runs | Correct iterative algorithm and code-only output |
| Hard Algorithm | exact prompt + 4 recorded runs | 5,000-case harness; O(log n) partition and correct edge cases |
Run dates: budget suite 2026-06-16T20:31:30.728Z; premium suite 2026-06-21T00:00:00.000Z. Results are not a claim about every repository or prompt.
Two-test rubric, failure analysis, and task-shaped ranking
| Model | Accuracy | Latency | Output tokens | Run cost | Failure / qualification note |
|---|---|---|---|---|---|
| GPT-5.4 Pro | 100/100 | 206116 ms | 548 | $0.103 | Passes all 5,000+ randomized cases and every edge case; genuine O(log(min(m,n))) partition search, returns a float, code-only, and explicitly raises on two empty lists. Correct — but the slowest run by far (over three minutes), and now that real usage is reported, comfortably the most expensive. |
| Claude Opus 4.8 | 100/100 | 3790 ms | 382 | $0.011 | Passes all 5,000+ randomized cases and every edge case; genuine O(log(min(m,n))) partition, float return, code-only. The fastest correct solution in this task. |
| GLM 5.2 (Max) | 100/100 | 47713 ms | 2148 | $0.010 | Passes all 5,000+ randomized cases and every edge case with a genuine O(log) partition and float return. The <think> block ahead of the code is GLM's reasoning channel surfaced by our gateway, not reasoning dumped into the answer — GLM's actual content is the clean code block — so it scores level with the other correct solutions, as the cheapest of them. |
| Gemini 3.1 Pro | 99/100 | 17413 ms | 320 | $0.004 | Passes all tests with a clean, minimal O(log) partition and float return, code-only. Docked one point only because two empty lists yield NaN rather than an explicit guard (not required by the prompt). |
Availability caveat: short code tests do not establish repository-scale debugging, multi-file tool use, or agent reliability. The fastest acceptable verdict must therefore clear the correctness rule before speed is considered.
Task-shaped cost ranking (20,000 tasks/month)
| Rank | Model | Effective monthly | Measured verbosity |
|---|---|---|---|
| 1 | Amazon Nova Micro | $18.26 | 0.76× |
| 2 | Amazon Nova Lite | $32.74 | 0.91× |
| 3 | GPT-5 Nano | $36.00 | Unavailable; neutral fallback |
| 4 | Gemini 2.5 Flash Lite | $56.00 | Unavailable; neutral fallback |
| 5 | GPT-OSS 20B | $65.16 | 2.93× |
| 6 | Ministral 8B | $65.58 | 0.93× |
Verified 2026-08-08. full prompt/run evidence →
Try these models for Translation →Volume, requirement-gate, and pricing cross-check for Translation
Monthly spend ladder at Translation batch shape
| Calls / month | Overall pick monthly | Budget pick monthly | Overall − budget delta |
|---|---|---|---|
| 50,000 | $15.00 | $15.00 | $0.0000 |
| 100,000 | $30.00 | $30.00 | $0.0000 |
| 200,000 | $60.00 | $60.00 | $0.0000 |
| 500,000 | $150.00 | $150.00 | $0.0000 |
Monthly cost = task price/M × (1,000 input + 1,000 output tokens) × calls ÷ 1,000,000, at 0.5×, 1×, 2×, 5× the published 100,000-call/month baseline.
Requirement-gate margin for the picked models
| Model | Requirement | Measured value | Margin / result |
|---|---|---|---|
| — | No hard requirement filter for this task | requires: {} | All qualified candidates pass by default |
Requirement gates are hard filters, not down-ranking: a model failing any row here is excluded from Translation candidates entirely, regardless of price or speed.
Pricing-page cross-link for each pick
| Model | Task-weighted $/M | Monthly at published volume | Measured throughput |
|---|---|---|---|
| Muse Spark 1.3 Contributor | $0.15 | $30.00 | Unavailable |
| GPT-OSS 120B (Cerebras) | $0.55 | $110.00 | 2450 tok/s |
| Gemini 3.1 Pro | $7.00 | $1400.00 | 55 tok/s |
Evidence coverage: 0 of 49 candidates have a graded run. No graded accuracy evidence exists for this task; the ranking above is a requirements-and-price match, not a quality claim.
Test the Translation picks side by side →Verified 2026-08-08. "Unavailable" means no compatible dated evidence was found for that field; it is never treated as zero. Dated source · Full ranking and rubric
Batch 10 translation economics
1. Token-density sensitivity per million source words
| Token density | Source volume | Expansion ratio | Billing tokens | API cost / 1M source words |
|---|---|---|---|---|
| 1.0 tokens/word | 1,000,000 source words | 1.25× output expansion | 1.00M input + 1.25M output tokens | $0.35 |
| 1.3 tokens/word | 1,000,000 source words | 1.25× output expansion | 1.30M input + 1.63M output tokens | $0.46 |
| 1.8 tokens/word | 1,000,000 source words | 1.25× output expansion | 1.80M input + 2.25M output tokens | $0.63 |
The explicit User-supplied token-density range is 1.0, 1.3, and 1.8 tokens per source word. Output expansion is held at 1.25× for comparison.
2. Glossary overhead sensitivity
| Reusable prompt tokens | Cost / 1K-word call | Cache treatment | Input status |
|---|---|---|---|
| 0 | $0.0003 | Cache rate: Unavailable unless provider-specific | Reusable terminology: User-supplied |
| 1,000 | $0.0004 | Cache rate: Unavailable unless provider-specific | Reusable terminology: User-supplied |
| 5,000 | $0.0008 | Cache rate: Unavailable unless provider-specific | Reusable terminology: User-supplied |
| 20,000 | $0.0023 | Cache rate: Unavailable unless provider-specific | Reusable terminology: User-supplied |
Missing cache pricing is Unavailable, never zero. The 1,000 input and 1,000 output base is the published task shape.
3. Human post-edit crossover
| Review minutes / 1K words | Hourly rate | API + review / 1K words | Quality evidence |
|---|---|---|---|
| 0 | $30/hour User-supplied | $0.0003 | Language-pair quality: Unavailable |
| 2 | $30/hour User-supplied | $1.00 | Language-pair quality: Unavailable |
| 5 | $30/hour User-supplied | $2.50 | Language-pair quality: Unavailable |
| 10 | $30/hour User-supplied | $5.00 | Language-pair quality: Unavailable |
Stale-data behavior: dated registry values are snapshots. If a source is older than the page verification date, or a provider changes its policy/pricing/model, re-verify before production use; unknown values remain Unavailable.
Verified 2026-08-08. Luna is the data owner. “Unavailable” means no compatible dated evidence was found; it is never zero, an estimate, or a guessed policy. Stale-data behavior: dated registry values are snapshots. If a source is older than the page verification date, or a provider changes its policy/pricing/model, re-verify before production use; unknown values remain Unavailable. Dated registry source · Run this scenario yourself →
Which models rank highest for Translation?
"Fit" is a requirements match, not a quality benchmark — it combines price, measured speed, context window, and (where we have run it) graded accuracy on this task. Formula below.
| # | Model | Provider | Fit | Evidence | Task price/M | Tokens/sec | Context | Scored on |
|---|---|---|---|---|---|---|---|---|
| 1 | Muse Spark 1.3 Contributor | Meta | 88 | — | $0.15 | — | 1.0M | price, context |
| 2 | Gemini 2.5 Flash LiteLegacy | 81 | — | $0.25 | — | 1M | price, context | |
| 3 | GPT-OSS 120B (Cerebras) | Cerebras | 70 | — | $0.55 | 2450 | 131K | price, context, speed |
| 4 | Gemini 2.5 FlashLegacy | 60 | — | $1.40 | — | 1M | price, context | |
| 5 | GPT-OSS 20B | Groq | 60 | — | $0.19 | 1120 | 131K | price, context, speed |
| 6 | GPT-4o MiniLegacy | OpenAI | 59 | — | $0.38 | — | 128K | price, context |
| 7 | Gemini 3.7 Flash | 55 | — | $2.25 | — | 1.0M | price, context | |
| 8 | GLM 4.7 (Cerebras) | Cerebras | 54 | — | $2.50 | 1980 | 200K | price, context, speed |
What will Translation cost?
At 100,000 translation batch calls/month:
| Model | Task price/M | Est. monthly cost |
|---|---|---|
| Muse Spark 1.3 Contributor | $0.15 | $30.00 |
| Gemini 2.5 Flash Lite | $0.25 | $50.00 |
| GPT-OSS 120B (Cerebras) | $0.55 | $110.00 |
How is the best LLM for Translation ranked?
Weights: evidence 0%, price 50%, speed 35%, context 15%.
Requirements: none — every current model is eligible. 49 models eligible.
Price and context sub-scores are min-max normalised (log-scaled) within this task's eligible set only. Speed uses measured tokens/sec only — estimated rows are excluded. A model missing a measurement is never scored as zero: its weight is redistributed across the components we do have, and "Scored on" in the table above shows exactly which ones.
Prices verified 2026-08-08.
What related resources help with Translation?
Where can you find evidence and costs for Translation?
What are common questions about the best LLM for Translation?
Do all models translate equally well?
Quality varies meaningfully by language pair and is not something our current evidence suite measures — test your specific language pairs against the top picks before committing.
Is a reasoning model better for translation?
Rarely — translation is closer to pattern transfer than multi-step reasoning, so the extra latency and cost of a reasoning mode usually isn't worth it.
Does context window matter for translation?
Only for very long documents translated in one call — most translation requests are short enough that any current model's window is more than sufficient.