Mistral Large 4 High vs Qwen3.8 27B Xhigh: which is faster?

This comparison puts Mistral's Mistral Large 4 (High) next to Alibaba's Qwen3.8 27B (Xhigh). The biggest gap is output speed, where Mistral Large 4 High records 139 tokens per second and Qwen3.8 27B Xhigh 38 tokens per second. Everything below is read from the benchmark, price and speed data in Franklin AI's catalogue; nothing here comes from a hands-on test.

How fast they respond

Once a response is under way, Mistral Large 4 High produces roughly 139 tokens per second against 38 tokens per second for Qwen3.8 27B Xhigh. Mistral Large 4 High is ahead by 264% on throughput. The wait for the first token averages 1.7 seconds on Mistral Large 4 High against 4.1 seconds on Qwen3.8 27B Xhigh. Mistral Large 4 High gets to its first token 2.3 times as fast.

Put together, a 2,000-token answer would take about 16 seconds from Mistral Large 4 High and 56 seconds from Qwen3.8 27B Xhigh, counting the first-token wait.

Benchmark scores

The catalogue puts Mistral Large 4 High at 26.1 and Qwen3.8 27B Xhigh at 20.1 on its intelligence index. A lead of this size is hard to dismiss as noise. Put another way, Qwen3.8 27B Xhigh would need 30% more on this index to draw level.

For context, that places Mistral Large 4 High 65th and Qwen3.8 27B Xhigh 78th out of 109 scored models still on offer.

For coding or maths work specifically, the catalogue has no dedicated index on either side yet.

Pricing depends on your mix

The list prices are $0.70 in and $2.10 out for Mistral Large 4 High, and $0.50 in and $3.00 out for Qwen3.8 27B Xhigh, per million tokens. The rates cross over. Qwen3.8 27B Xhigh is cheaper to feed and Mistral Large 4 High is cheaper to hear back from.

A sample workload of 1 million input tokens and 200,000 output tokens costs about the same on either: $1.12 against $1.10. A bigger example, writing 2,000 drafts of 2,000 output tokens from 300-token prompts, costs $8.82 on Mistral Large 4 High and $12.30 on Qwen3.8 27B Xhigh: a gap of $3.48. Blended at three input tokens per output token, the effective rates are $1.05 for Mistral Large 4 High and $1.125 for Qwen3.8 27B Xhigh.

Release timing

Mistral Large 4 High arrived on October 29, 2026, about 2 months after Qwen3.8 27B Xhigh (August 14, 2026).

Value for money

A blunt value measure, blended price over intelligence score, puts Mistral Large 4 High at $0.040 and Qwen3.8 27B Xhigh at $0.056 per point. Mistral Large 4 High scores higher and costs less per point, so on price per point there is nothing to trade off.