GPT-5.6 Sol Ultrafast on Cerebras CS-4: 750 output tokens/sec

Vendor benchmark, single-stream decode, GPT-5.6 Sol Ultrafast mode on Cerebras CS-4. No batch size or concurrency specified. Comparison: 11x faster than Fable 5, 5x faster than Opus 4.8 Fast mode.

Vendor claim from Cerebras blog (Aug 13, 2026). This is a single-stream decode throughput figure for a specific model (GPT-5.6 Sol) in Ultrafast mode. No independent third-party verification. The figure is for the CS-4 system with 3 WSE-3 Turbo wafers. Cerebras publishes no standard benchmark (MLPerf) results. LOAD-CONDITION AUDIT (bq- 2026-10-04). SCENARIO: NOT STATED by the source - absent, not inferred. A throughput figure with no scenario cannot be placed against an MLPerf Offline or Server number (they are different measurements; on identical hardware the Offline-minus-Server difference is negative in 5 of 9 systems measured). ABSENT CONDITIONS (3 of 7 checked: model, hardware, batch-or-concurrency, input length, output length, engine version, warmup): - engine version: unrecoverable from alt-cerebras-ultrafast - engine_id=alt-cerebras-ultrafast is named but no build/commit/container digest is published - input length: unrecoverable from alt-cerebras-ultrafast - would need the run's ISL or the prepared dataset's input-length profile - warmup: unrecoverable from alt-cerebras-ultrafast - would need the raw harness log's warmup phase or the benchmark command line (loadgen min_duration/min_query_count, or trtllm-bench --warm_up N / --num_runs N) None of these were filled in. A plausible value is worse than an absent one: it would be trusted.